## The Kidekin TRNG Hardware Random Number Generator

Yesterday, I received my Kidekin TRNG hardware random number generator. I was eager to purchase this, because on the Tindie website, the first 2 people to purchase the RNG would get $50 off, making the device$30 total. I quickly ordered one. Hilariously enough, I received a letter from the supplier that I was their first customer! Hah!

Upon opening the package, I noticed the size of the TRNG. It's roughly 10.5 cm from end-to-end which makes it somewhat awkward for a device sitting in your USB port on your laptop. It would work fine sitting in the back of a desktop or server, out of the way, but on my Thinkpad T61, it's a bit large to be sitting there 24/7 feeding my kernel CSPRNG.

Plugging the device in, the kernel actually sees two USB devices, not just one, and sets them up as /dev/ttyUSB0 and /dev/ttyUSB1. Curious. Downloading the software ZIP file from their webpage, and looking through it, the following UDEV rules are provided:

$cat /etc/udev/rules.d/98-kidekin.rules #SYMLINK+= method works on more systems, if it does not on your system, please switch to the NAME= method. #disable the unused port. #SUBSYSTEM=="tty", ATTRS{interface}=="kidekin_trng", ATTRS{bInterfaceNumber}=="00", NAME="kidekin_dont_use", MODE="0000", ENV{ID_MM_DEVICE_IGNORE}="1", ENV{ID_MM_CANDIDATE}="0" SUBSYSTEM=="tty", ATTRS{interface}=="kidekin_trng", ATTRS{bInterfaceNumber}=="00", SYMLINK+="kidekin_dont_use", MODE="0000", ENV{ID_MM_DEVICE_IGNORE}="1", ENV{ID_MM_CANDIDATE}="0" #connect kidekin TRNG to /dev/random #SUBSYSTEM=="tty", ATTRS{interface}=="kidekin_trng", ATTRS{bInterfaceNumber}=="01", NAME="kidekin_trng", MODE="0777", RUN+="/bin/stty raw -echo -crtscts -F /dev/kidekin_trng speed 3000000", ENV{ID_MM_DEVICE_IGNORE}="1", ENV{ID_MM_CANDIDATE}="0" SUBSYSTEM=="tty", ATTRS{interface}=="kidekin_trng", ATTRS{bInterfaceNumber}=="01", SYMLINK+="kidekin_trng", MODE="0777", RUN+="/bin/stty raw -echo -crtscts -F /dev/kidekin_trng speed 3000000", ENV{ID_MM_DEVICE_IGNORE}="1", ENV{ID_MM_CANDIDATE}="0" SUBSYSTEM=="tty", ATTRS{interface}=="kidekin_trng", ATTRS{bInterfaceNumber}=="01", RUN+="/etc/init.d/rng-tools restart" This is a bit assuming, and a bit overdoing it IMO, so I simplified it, and setup the following: SUBSYSTEM=="tty", ATTRS{interface}=="kidekin_trng", ATTRS{bInterfaceNumber}=="01", SYMLINK+="kidekin", MODE="0777", RUN+="/bin/stty raw -echo -crtscts -F /dev/kidekin speed 3000000", ENV{ID_MM_DEVICE_IGNORE}="1", ENV{ID_MM_CANDIDATE}="0" This avoids setting up a "do not use" symlink for the unnecessary USB device, and changes the symlink of the usable USB device to /dev/kidekin. This also doesn't restart rngd(8), as I'll administer that on my own. At this point, I am ready for testing. First and foremost, I wanted to test its throughput: $ dd if=/dev/kidekin count=1G | pv -a > /dev/null
[ 282KiB/s]

The device held stable at 282 KBps or roughly 2.2 Mbps. This is 75.2 KBps per dollar for my $30 purchase. Not bad. The Kidekin is based on astable free running oscillators, or multivibrators. Unfortunately, a security proof does not accompany the device. So, while this may hold up to the suite of randomness tests, the output may not be cryptographically secure, and could also potentially be backdoored, as verifying the hardware is not easily doable. So, let's see if it at least holds up to the randomness tests. I created a 256 MB file, and ran the standard suites of tests: $ dd if=/dev/kidekin of=entropy.kidekin bs=1M count=256 iflag=fullblock
256+0 records in
256+0 records out
268435456 bytes (268 MB) copied, 928.326 s, 289 kB/s

At this point, I can start my testing. First, let's quantify the amount of entropy per byte, as well as some basic tests with ent(1):

$ent entropy.kidekin Entropy = 7.999999 bits per byte. Optimum compression would reduce the size of this 268435456 byte file by 0 percent. Chi square distribution for 268435456 samples is 248.92, and randomly would exceed this value 59.56 percent of the times. Arithmetic mean value of data bytes is 127.4924 (127.5 = random). Monte Carlo value for Pi is 3.141825693 (error 0.01 percent). Serial correlation coefficient is -0.000003 (totally uncorrelated = 0.0). Everything good so far. How about the FIPS 140-2 tests for randomness: $ rngtest < entropy.kidekin
rngtest 2-unofficial-mt.14
Copyright (c) 2004 by Henrique de Moraes Holschuh
This is free software; see the source for copying conditions.  There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

rngtest: starting FIPS tests...
rngtest: entropy source exhausted!
rngtest: bits received from input: 2147483648
rngtest: FIPS 140-2 successes: 107292
rngtest: FIPS 140-2 failures: 82
rngtest: FIPS 140-2(2001-10-10) Monobit: 14
rngtest: FIPS 140-2(2001-10-10) Poker: 13
rngtest: FIPS 140-2(2001-10-10) Runs: 26
rngtest: FIPS 140-2(2001-10-10) Long run: 30
rngtest: FIPS 140-2(2001-10-10) Continuous run: 0
rngtest: input channel speed: (min=317.891; avg=7386.982; max=19073.486)Mibits/s
rngtest: FIPS tests speed: (min=6.563; avg=109.376; max=114.901)Mibits/s
rngtest: Program run time: 19261018 microseconds
$echo$?
1

Again, so far so good. Some failures are expected with random input of this size. 82 failures versus 107292 successes is right on par with the tests. Now the Dieharder battery of tests:

$dieharder -a < entropy.kidekin #=============================================================================# # dieharder version 3.31.1 Copyright 2003 Robert G. Brown # #=============================================================================# rng_name |rands/second| Seed | mt19937| 8.99e+07 | 722892634| #=============================================================================# test_name |ntup| tsamples |psamples| p-value |Assessment #=============================================================================# diehard_birthdays| 0| 100| 100|0.87388974| PASSED diehard_operm5| 0| 1000000| 100|0.25081726| PASSED diehard_rank_32x32| 0| 40000| 100|0.80329585| PASSED diehard_rank_6x8| 0| 100000| 100|0.87234234| PASSED diehard_bitstream| 0| 2097152| 100|0.27873738| PASSED diehard_opso| 0| 2097152| 100|0.05958924| PASSED diehard_oqso| 0| 2097152| 100|0.10540020| PASSED diehard_dna| 0| 2097152| 100|0.30006047| PASSED diehard_count_1s_str| 0| 256000| 100|0.43809130| PASSED diehard_count_1s_byt| 0| 256000| 100|0.29758303| PASSED diehard_parking_lot| 0| 12000| 100|0.78081639| PASSED diehard_2dsphere| 2| 8000| 100|0.58294587| PASSED diehard_3dsphere| 3| 4000| 100|0.04012616| PASSED diehard_squeeze| 0| 100000| 100|0.97651988| PASSED diehard_sums| 0| 100| 100|0.01875349| PASSED diehard_runs| 0| 100000| 100|0.17566659| PASSED diehard_runs| 0| 100000| 100|0.78887310| PASSED diehard_craps| 0| 200000| 100|0.16369886| PASSED diehard_craps| 0| 200000| 100|0.42148915| PASSED marsaglia_tsang_gcd| 0| 10000000| 100|0.27534860| PASSED marsaglia_tsang_gcd| 0| 10000000| 100|0.45190499| PASSED sts_monobit| 1| 100000| 100|0.88204376| PASSED sts_runs| 2| 100000| 100|0.15277754| PASSED sts_serial| 1| 100000| 100|0.71489026| PASSED sts_serial| 2| 100000| 100|0.85005457| PASSED sts_serial| 3| 100000| 100|0.77631916| PASSED sts_serial| 3| 100000| 100|0.81111751| PASSED sts_serial| 4| 100000| 100|0.72512842| PASSED sts_serial| 4| 100000| 100|0.68758000| PASSED sts_serial| 5| 100000| 100|0.69083583| PASSED sts_serial| 5| 100000| 100|0.09706031| PASSED sts_serial| 6| 100000| 100|0.52758972| PASSED sts_serial| 6| 100000| 100|0.27970465| PASSED sts_serial| 7| 100000| 100|0.07925569| PASSED sts_serial| 7| 100000| 100|0.25874891| PASSED sts_serial| 8| 100000| 100|0.33647659| PASSED sts_serial| 8| 100000| 100|0.80952471| PASSED sts_serial| 9| 100000| 100|0.99948911| WEAK sts_serial| 9| 100000| 100|0.32461849| PASSED sts_serial| 10| 100000| 100|0.69360795| PASSED sts_serial| 10| 100000| 100|0.96022345| PASSED sts_serial| 11| 100000| 100|0.91349333| PASSED sts_serial| 11| 100000| 100|0.95918606| PASSED sts_serial| 12| 100000| 100|0.69821905| PASSED sts_serial| 12| 100000| 100|0.57652285| PASSED sts_serial| 13| 100000| 100|0.28393582| PASSED sts_serial| 13| 100000| 100|0.45849491| PASSED sts_serial| 14| 100000| 100|0.30832853| PASSED sts_serial| 14| 100000| 100|0.89099315| PASSED sts_serial| 15| 100000| 100|0.87022105| PASSED sts_serial| 15| 100000| 100|0.06938123| PASSED sts_serial| 16| 100000| 100|0.79568629| PASSED sts_serial| 16| 100000| 100|0.53218489| PASSED rgb_bitdist| 1| 100000| 100|0.38552808| PASSED rgb_bitdist| 2| 100000| 100|0.79403454| PASSED rgb_bitdist| 3| 100000| 100|0.66811643| PASSED rgb_bitdist| 4| 100000| 100|0.84954470| PASSED rgb_bitdist| 5| 100000| 100|0.90198903| PASSED rgb_bitdist| 6| 100000| 100|0.98808244| PASSED rgb_bitdist| 7| 100000| 100|0.25730860| PASSED rgb_bitdist| 8| 100000| 100|0.43237015| PASSED rgb_bitdist| 9| 100000| 100|0.90916135| PASSED rgb_bitdist| 10| 100000| 100|0.81131338| PASSED rgb_bitdist| 11| 100000| 100|0.31361128| PASSED rgb_bitdist| 12| 100000| 100|0.40786889| PASSED rgb_minimum_distance| 2| 10000| 1000|0.03358258| PASSED rgb_minimum_distance| 3| 10000| 1000|0.99298827| PASSED rgb_minimum_distance| 4| 10000| 1000|0.47721533| PASSED rgb_minimum_distance| 5| 10000| 1000|0.86641982| PASSED rgb_permutations| 2| 100000| 100|0.10084049| PASSED rgb_permutations| 3| 100000| 100|0.99560585| WEAK rgb_permutations| 4| 100000| 100|0.42217190| PASSED rgb_permutations| 5| 100000| 100|0.95466090| PASSED rgb_lagged_sum| 0| 1000000| 100|0.64120688| PASSED rgb_lagged_sum| 1| 1000000| 100|0.22106106| PASSED rgb_lagged_sum| 2| 1000000| 100|0.41244281| PASSED rgb_lagged_sum| 3| 1000000| 100|0.98880097| PASSED rgb_lagged_sum| 4| 1000000| 100|0.78380177| PASSED rgb_lagged_sum| 5| 1000000| 100|0.25533777| PASSED rgb_lagged_sum| 6| 1000000| 100|0.78150371| PASSED rgb_lagged_sum| 7| 1000000| 100|0.53903267| PASSED rgb_lagged_sum| 8| 1000000| 100|0.04436257| PASSED rgb_lagged_sum| 9| 1000000| 100|0.77174302| PASSED rgb_lagged_sum| 10| 1000000| 100|0.54862612| PASSED rgb_lagged_sum| 11| 1000000| 100|0.48691334| PASSED rgb_lagged_sum| 12| 1000000| 100|0.06308057| PASSED rgb_lagged_sum| 13| 1000000| 100|0.42530804| PASSED rgb_lagged_sum| 14| 1000000| 100|0.86907366| PASSED rgb_lagged_sum| 15| 1000000| 100|0.66262930| PASSED rgb_lagged_sum| 16| 1000000| 100|0.85485044| PASSED rgb_lagged_sum| 17| 1000000| 100|0.39817394| PASSED rgb_lagged_sum| 18| 1000000| 100|0.90608610| PASSED rgb_lagged_sum| 19| 1000000| 100|0.94996515| PASSED rgb_lagged_sum| 20| 1000000| 100|0.78715690| PASSED rgb_lagged_sum| 21| 1000000| 100|0.93364519| PASSED rgb_lagged_sum| 22| 1000000| 100|0.84438533| PASSED rgb_lagged_sum| 23| 1000000| 100|0.77439531| PASSED rgb_lagged_sum| 24| 1000000| 100|0.12530311| PASSED rgb_lagged_sum| 25| 1000000| 100|0.79035917| PASSED rgb_lagged_sum| 26| 1000000| 100|0.93286961| PASSED rgb_lagged_sum| 27| 1000000| 100|0.32567247| PASSED rgb_lagged_sum| 28| 1000000| 100|0.39563718| PASSED rgb_lagged_sum| 29| 1000000| 100|0.15628693| PASSED rgb_lagged_sum| 30| 1000000| 100|0.69368810| PASSED rgb_lagged_sum| 31| 1000000| 100|0.00197963| WEAK rgb_lagged_sum| 32| 1000000| 100|0.23325783| PASSED rgb_kstest_test| 0| 10000| 1000|0.18940877| PASSED dab_bytedistrib| 0| 51200000| 1|0.57007834| PASSED dab_dct| 256| 50000| 1|0.76567665| PASSED Preparing to run test 207. ntuple = 0 dab_filltree| 32| 15000000| 1|0.60537852| PASSED dab_filltree| 32| 15000000| 1|0.78894908| PASSED Preparing to run test 208. ntuple = 0 dab_filltree2| 0| 5000000| 1|0.11775507| PASSED dab_filltree2| 1| 5000000| 1|0.34799105| PASSED Preparing to run test 209. ntuple = 0 dab_monobit2| 12| 65000000| 1|0.69182598| PASSED  Finally, a visual check on the data, even though it's safe to assume that it's "true random" given the previous testing: $ dd if=white.bmp of=entropy.kidekin bs=1 count=54 conv=notrunc
54+0 records in
54+0 records out
54 bytes (54 B) copied, 0.000547208 s, 98.7 kB/s
$gimp entropy.kidekin # convert to grayscale, export as "entropy.png"$ optipng entropy.png
** Processing: entropy.png
512x512 pixels, 8 bits/pixel, grayscale
Input IDAT size = 250107 bytes
Input file size = 250564 bytes

Trying:
zc = 9  zm = 8  zs = 0  f = 0		IDAT size = 215319
zc = 9  zm = 8  zs = 1  f = 0		IDAT size = 214467
zc = 1  zm = 8  zs = 2  f = 0		IDAT size = 214467
zc = 9  zm = 8  zs = 3  f = 0		IDAT size = 214467

Selecting parameters:
zc = 1  zm = 8  zs = 2  f = 0		IDAT size = 214467

Output IDAT size = 214467 bytes (35640 bytes decrease)
Output file size = 214564 bytes (36000 bytes = 14.37% decrease)

And the result is:

My conclusion of the Kidekin TRNG is positive. I love the throughput of the device, loved the price, and aside from the UDEV rule, it is plug-and-play. Unfortunately, the TRNG is a bit on the big side for a physical device, and because it doesn't come with a security proof, and the hardware design is closed, I would be skeptical to trust it for your random numbers directly. Instead, I would recommend adding it the Linux kernel's CSPRNG, and rely on /dev/urandom instead. This is trivial with rngd(8). But, overall, I am very pleased with the device, and which I had actually purchased a second one.

## Additional Testing Of The rtl-sdr Dongle As A HWRNG

A couple days ago, I put up a post about using the Realtek SDR dongles as a hardware true random number generator. I only tested the randomness of a 512 MB file. I thought this time, I would but a bit more stock into it. In this case, I let it run for a while, until it was 1.8 GB in size. Interestingly enough, it stopped getting bigger after that point. Not sure why. However, I ran more tests on that 1.8 GB file. Creating this file took its time:

$tail -f /run/rtl_entropy.fifo | dd of=random.img iflag=fullblock 3554130+0 records in 3554130+0 records out 1819714560 bytes (1.8 GB) copied, 3897.22 s, 467 kB/s  This filled up a bit faster than I had previously tested, going at a clip of about 3.826 Mbps. Now it was time for the testing: $ ent random.img
Entropy = 8.000000 bits per byte.

Optimum compression would reduce the size
of this 1819714560 byte file by 0 percent.

Chi square distribution for 1819714560 samples is 246.86, and randomly
would exceed this value 63.11 percent of the times.

Arithmetic mean value of data bytes is 127.4990 (127.5 = random).
Monte Carlo value for Pi is 3.141611317 (error 0.00 percent).
Serial correlation coefficient is 0.000013 (totally uncorrelated = 0.0).

It passes with flying colors on entropy estimation, compression, chi-square distributions, arithmetic mean, the Monte Carlo estimation for Pi, and serial correlation. Testing further, I ran it through the FIPS 140-2 tests:

$rngtest < random.img rngtest 2-unofficial-mt.14 Copyright (c) 2004 by Henrique de Moraes Holschuh This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. rngtest: starting FIPS tests... rngtest: entropy source exhausted! rngtest: bits received from input: 14557716480 rngtest: FIPS 140-2 successes: 727288 rngtest: FIPS 140-2 failures: 597 rngtest: FIPS 140-2(2001-10-10) Monobit: 99 rngtest: FIPS 140-2(2001-10-10) Poker: 57 rngtest: FIPS 140-2(2001-10-10) Runs: 210 rngtest: FIPS 140-2(2001-10-10) Long run: 233 rngtest: FIPS 140-2(2001-10-10) Continuous run: 0 rngtest: input channel speed: (min=114.212; avg=6626.942; max=9536.743)Mibits/s rngtest: FIPS tests speed: (min=61.133; avg=147.877; max=151.377)Mibits/s rngtest: Program run time: 96034230 microseconds You have new mail.$ echo $? 1 Finally, the beast of beasts, I ran it through every Dieharder test. This took some time to complete. Here is a listing of the tests that it went through: $ dieharder -l
#=============================================================================#
#            dieharder version 3.31.1 Copyright 2003 Robert G. Brown          #
#=============================================================================#
Installed dieharder tests:
Test Number	                     Test Name	              Test Reliability
===============================================================================
-d 0  	                  Diehard Birthdays Test	      Good
-d 1  	                     Diehard OPERM5 Test	      Good
-d 2  	          Diehard 32x32 Binary Rank Test	      Good
-d 3  	            Diehard 6x8 Binary Rank Test	      Good
-d 4  	                  Diehard Bitstream Test	      Good
-d 5  	                            Diehard OPSO	   Suspect
-d 6  	                       Diehard OQSO Test	   Suspect
-d 7  	                        Diehard DNA Test	   Suspect
-d 8  	      Diehard Count the 1s (stream) Test	      Good
-d 9  	        Diehard Count the 1s Test (byte)	      Good
-d 10  	                Diehard Parking Lot Test	      Good
-d 11  	Diehard Minimum Distance (2d Circle) Test	      Good
-d 12  	Diehard 3d Sphere (Minimum Distance) Test	      Good
-d 13  	                    Diehard Squeeze Test	      Good
-d 14  	                       Diehard Sums Test	Do Not Use
-d 15  	                       Diehard Runs Test	      Good
-d 16  	                      Diehard Craps Test	      Good
-d 17  	            Marsaglia and Tsang GCD Test	      Good
-d 100  	                        STS Monobit Test	      Good
-d 101  	                           STS Runs Test	      Good
-d 102  	           STS Serial Test (Generalized)	      Good
-d 200  	               RGB Bit Distribution Test	      Good
-d 201  	   RGB Generalized Minimum Distance Test	      Good
-d 202  	                   RGB Permutations Test	      Good
-d 203  	                     RGB Lagged Sum Test	      Good
-d 204  	        RGB Kolmogorov-Smirnov Test Test	      Good
-d 205  	                       Byte Distribution	      Good
-d 206  	                                 DAB DCT	      Good
-d 207  	                      DAB Fill Tree Test	      Good
-d 208  	                    DAB Fill Tree 2 Test	      Good
-d 209  	                      DAB Monobit 2 Test	      Good

So here are the results:

 $dieharder -a < random.img #=============================================================================# # dieharder version 3.31.1 Copyright 2003 Robert G. Brown # #=============================================================================# rng_name |rands/second| Seed | mt19937| 1.25e+08 | 169223456| #=============================================================================# test_name |ntup| tsamples |psamples| p-value |Assessment #=============================================================================# diehard_birthdays| 0| 100| 100|0.91937112| PASSED diehard_operm5| 0| 1000000| 100|0.77213572| PASSED diehard_rank_32x32| 0| 40000| 100|0.04709503| PASSED diehard_rank_6x8| 0| 100000| 100|0.93031877| PASSED diehard_bitstream| 0| 2097152| 100|0.12183977| PASSED diehard_opso| 0| 2097152| 100|0.96023625| PASSED diehard_oqso| 0| 2097152| 100|0.61237304| PASSED diehard_dna| 0| 2097152| 100|0.66045974| PASSED diehard_count_1s_str| 0| 256000| 100|0.16999968| PASSED diehard_count_1s_byt| 0| 256000| 100|0.00992823| PASSED diehard_parking_lot| 0| 12000| 100|0.69592283| PASSED diehard_2dsphere| 2| 8000| 100|0.95358410| PASSED diehard_3dsphere| 3| 4000| 100|0.89028448| PASSED diehard_squeeze| 0| 100000| 100|0.81631204| PASSED diehard_sums| 0| 100| 100|0.03559934| PASSED diehard_runs| 0| 100000| 100|0.75027140| PASSED diehard_runs| 0| 100000| 100|0.43076351| PASSED diehard_craps| 0| 200000| 100|0.57749359| PASSED diehard_craps| 0| 200000| 100|0.00599436| PASSED marsaglia_tsang_gcd| 0| 10000000| 100|0.60121369| PASSED marsaglia_tsang_gcd| 0| 10000000| 100|0.04254338| PASSED sts_monobit| 1| 100000| 100|0.94352358| PASSED sts_runs| 2| 100000| 100|0.77549833| PASSED sts_serial| 1| 100000| 100|0.46198961| PASSED sts_serial| 2| 100000| 100|0.46002706| PASSED sts_serial| 3| 100000| 100|0.73076110| PASSED sts_serial| 3| 100000| 100|0.90967100| PASSED sts_serial| 4| 100000| 100|0.32002297| PASSED sts_serial| 4| 100000| 100|0.07478887| PASSED sts_serial| 5| 100000| 100|0.27486408| PASSED sts_serial| 5| 100000| 100|0.57409336| PASSED sts_serial| 6| 100000| 100|0.05095556| PASSED sts_serial| 6| 100000| 100|0.06341272| PASSED sts_serial| 7| 100000| 100|0.00941089| PASSED sts_serial| 7| 100000| 100|0.53679805| PASSED sts_serial| 8| 100000| 100|0.00122125| WEAK sts_serial| 8| 100000| 100|0.16239101| PASSED sts_serial| 9| 100000| 100|0.24007712| PASSED sts_serial| 9| 100000| 100|0.02659941| PASSED sts_serial| 10| 100000| 100|0.64616186| PASSED sts_serial| 10| 100000| 100|0.78783799| PASSED sts_serial| 11| 100000| 100|0.77618602| PASSED sts_serial| 11| 100000| 100|0.33875893| PASSED sts_serial| 12| 100000| 100|0.50423715| PASSED sts_serial| 12| 100000| 100|0.77528158| PASSED sts_serial| 13| 100000| 100|0.57625144| PASSED sts_serial| 13| 100000| 100|0.73422196| PASSED sts_serial| 14| 100000| 100|0.40891605| PASSED sts_serial| 14| 100000| 100|0.48542772| PASSED sts_serial| 15| 100000| 100|0.67319390| PASSED sts_serial| 15| 100000| 100|0.74730027| PASSED sts_serial| 16| 100000| 100|0.67519158| PASSED sts_serial| 16| 100000| 100|0.73171087| PASSED rgb_bitdist| 1| 100000| 100|0.87216594| PASSED rgb_bitdist| 2| 100000| 100|0.18831902| PASSED rgb_bitdist| 3| 100000| 100|0.16757216| PASSED rgb_bitdist| 4| 100000| 100|0.05327115| PASSED rgb_bitdist| 5| 100000| 100|0.75278396| PASSED rgb_bitdist| 6| 100000| 100|0.64749144| PASSED rgb_bitdist| 7| 100000| 100|0.20311557| PASSED rgb_bitdist| 8| 100000| 100|0.39994123| PASSED rgb_bitdist| 9| 100000| 100|0.52805289| PASSED rgb_bitdist| 10| 100000| 100|0.96091722| PASSED rgb_bitdist| 11| 100000| 100|0.97794399| PASSED rgb_bitdist| 12| 100000| 100|0.75009561| PASSED rgb_minimum_distance| 2| 10000| 1000|0.58923867| PASSED rgb_minimum_distance| 3| 10000| 1000|0.54294743| PASSED rgb_minimum_distance| 4| 10000| 1000|0.59446131| PASSED rgb_minimum_distance| 5| 10000| 1000|0.00047025| WEAK rgb_permutations| 2| 100000| 100|0.89040191| PASSED rgb_permutations| 3| 100000| 100|0.47917416| PASSED rgb_permutations| 4| 100000| 100|0.30964668| PASSED rgb_permutations| 5| 100000| 100|0.70217495| PASSED rgb_lagged_sum| 0| 1000000| 100|0.12796648| PASSED rgb_lagged_sum| 1| 1000000| 100|0.15077254| PASSED rgb_lagged_sum| 2| 1000000| 100|0.31141471| PASSED rgb_lagged_sum| 3| 1000000| 100|0.94974697| PASSED rgb_lagged_sum| 4| 1000000| 100|0.99256987| PASSED rgb_lagged_sum| 5| 1000000| 100|0.67854004| PASSED rgb_lagged_sum| 6| 1000000| 100|0.08600877| PASSED rgb_lagged_sum| 7| 1000000| 100|0.91633363| PASSED rgb_lagged_sum| 8| 1000000| 100|0.06794590| PASSED rgb_lagged_sum| 9| 1000000| 100|0.59024027| PASSED rgb_lagged_sum| 10| 1000000| 100|0.59285975| PASSED rgb_lagged_sum| 11| 1000000| 100|0.87178336| PASSED rgb_lagged_sum| 12| 1000000| 100|0.63401541| PASSED rgb_lagged_sum| 13| 1000000| 100|0.47202172| PASSED rgb_lagged_sum| 14| 1000000| 100|0.34616699| PASSED rgb_lagged_sum| 15| 1000000| 100|0.97221211| PASSED rgb_lagged_sum| 16| 1000000| 100|0.95576739| PASSED rgb_lagged_sum| 17| 1000000| 100|0.32367098| PASSED rgb_lagged_sum| 18| 1000000| 100|0.92792046| PASSED rgb_lagged_sum| 19| 1000000| 100|0.58128429| PASSED rgb_lagged_sum| 20| 1000000| 100|0.78197001| PASSED rgb_lagged_sum| 21| 1000000| 100|0.86068846| PASSED rgb_lagged_sum| 22| 1000000| 100|0.22496908| PASSED rgb_lagged_sum| 23| 1000000| 100|0.52387665| PASSED rgb_lagged_sum| 24| 1000000| 100|0.52748770| PASSED rgb_lagged_sum| 25| 1000000| 100|0.96442902| PASSED rgb_lagged_sum| 26| 1000000| 100|0.51298847| PASSED rgb_lagged_sum| 27| 1000000| 100|0.99123470| PASSED rgb_lagged_sum| 28| 1000000| 100|0.69774674| PASSED rgb_lagged_sum| 29| 1000000| 100|0.83646714| PASSED rgb_lagged_sum| 30| 1000000| 100|0.98573851| PASSED rgb_lagged_sum| 31| 1000000| 100|0.23580471| PASSED rgb_lagged_sum| 32| 1000000| 100|0.19150884| PASSED rgb_kstest_test| 0| 10000| 1000|0.67771558| PASSED dab_bytedistrib| 0| 51200000| 1|0.07152541| PASSED dab_dct| 256| 50000| 1|0.53841656| PASSED Preparing to run test 207. ntuple = 0 dab_filltree| 32| 15000000| 1|0.09092747| PASSED dab_filltree| 32| 15000000| 1|0.83382174| PASSED Preparing to run test 208. ntuple = 0 dab_filltree2| 0| 5000000| 1|0.37363586| PASSED dab_filltree2| 1| 5000000| 1|0.26890999| PASSED Preparing to run test 209. ntuple = 0 dab_monobit2| 12| 65000000| 1|0.80810458| PASSED  I don't have an image to look at to visually verify that there are no obvious patterns. At 1.8 GB, I feel that it would be just a bit too unwieldy anyway. So, I'll need to trust the previous tests for randomness that the data really is random. After these 3 series of tests, I can only conclude that using a Realtek SDR as a HWRNG will generate as "true random" data as you can hope for. ## Hardware RNG Through an rtl-sdr Dongle An rtl-sdr dongle allows you to receive radio frequency signals to your computer through a software interface. You can listen to Amateur Radio, watch analog television, listen to FM radio broadcasts, and a number of other things. I have a friend to uses it to monitor power usage at his house. However, I have a different use- true random number generation. The theory behind the RNG is by taking advantage of radio frequency noise such as atmospheric noise. which is caused by natural occurrences, such as weak galactic radiation from the center of our Milky Way Galaxy to the stronger local and remote lightning strikes. It's estimated that roughly 40 lightning strikes are hitting the Earth every second, which equates to about 3.5 million strikes per 24 hour period. Interestingly enough, this provides a great deal of entropy for a random number generator. Check out Blitzortung. It is a community run site, where volunteers can setup lightning monitoring stations and submit data to the server. Of course, it isn't an accurate picture of the entire globe, but you can at least get some idea of the scope of lightning strikes around the continents. Unfortunately, however, the rtl-sdr dongle won't get down to the frequencies necessary for sampling atmospheric noise; about 100 KHz to 10 MHz, and above 10 GHz. However, it can sample cosmic noise, man-made (urban and suburban) noise, solar noise, thermal noise, and other terrestrial noises that are well within the tuner frequency range of the dongle. In order to take advantage of this, you obviously need an rtl-sdr dongle. They're quite cheap, about$15 or so, and plug in via USB with an external antenna. Of course, the larger the antenna, the more terrestrial noise you'll be able to observe. With a standard telescoping antenna, I can observe about 3 Mbps of true random data.

The other piece, however, will be compiling and installing the rtl-entropy software. This will provide a FIFO file for observing the random data. Reading the random data can be done as you would read any regular file:

$sudo rtl_entropy -b -f 74M$ tail -f /run/rtl_entropy.fifo | dd of=/dev/null
^C8999+10 records in
9004+0 records out
4610048 bytes (4.6 MB) copied, 13.294 s, 347 kB/s

That's roughly 2.8 Mbps. Not bad for $15. Notice, that I passed the "-b" switch to detach the PID from the controlling TTY and background. Further, I am not tuning to the default frequency of 70 MHz, which is part of Band I in the North America band plan for television broadcasting. Instead, I am tuning to 74 MHz, which is in the middle of a break in the band plan, where no television broadcasting should be transmitted. Of course, you'll need to make sure you are tuning to a frequency that is less likely to encounter malicious interference. Even though the rtl_entropy daemon has built-in debiasing and FIPS randomness testing, a malicious source could interrupt with the operation of the output by transmitting on the frequency that you are listening to. In order to guarantee that you have random data, you should send it through a battery of standardized tests for randomness. One popular test for randomness are the FIPS 140-2 tests. Suppose I create a 512 MB file from my sdr-rtl dongle, I can test it as follows: $ rngtest < random.img
rngtest 2-unofficial-mt.14
Copyright (c) 2004 by Henrique de Moraes Holschuh
This is free software; see the source for copying conditions.  There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

rngtest: starting FIPS tests...
rngtest: entropy source exhausted!
rngtest: bits received from input: 83886080
rngtest: FIPS 140-2 successes: 4190
rngtest: FIPS 140-2 failures: 4
rngtest: FIPS 140-2(2001-10-10) Monobit: 0
rngtest: FIPS 140-2(2001-10-10) Poker: 1
rngtest: FIPS 140-2(2001-10-10) Runs: 1
rngtest: FIPS 140-2(2001-10-10) Long run: 2
rngtest: FIPS 140-2(2001-10-10) Continuous run: 0
rngtest: input channel speed: (min=174.986; avg=4379.165; max=4768.372)Mibits/s
rngtest: FIPS tests speed: (min=113.533; avg=147.777; max=150.185)Mibits/s
rngtest: Program run time: 560095 microseconds

It's expected to see some failures, but they should be outliers. There is also the Dieharder battery of randomness tests. This will take substantially longer to work through, but it can be done. Here are the first few lines:

$dieharder -a < random.img #=============================================================================# # dieharder version 3.31.1 Copyright 2003 Robert G. Brown # #=============================================================================# rng_name |rands/second| Seed | mt19937| 1.30e+08 | 334923062| #=============================================================================# test_name |ntup| tsamples |psamples| p-value |Assessment #=============================================================================# diehard_birthdays| 0| 100| 100|0.98331589| PASSED diehard_operm5| 0| 1000000| 100|0.12201131| PASSED diehard_rank_32x32| 0| 40000| 100|0.69993313| PASSED diehard_rank_6x8| 0| 100000| 100|0.55365877| PASSED diehard_bitstream| 0| 2097152| 100|0.85077208| PASSED diehard_opso| 0| 2097152| 100|0.76171650| PASSED  The whole dieharder results of my 512 MB random file can be found here. Last, but not least, it helps to observe the data visually. In this image, I created a plain white file in Gimp, that was 600x600 pixels in size. I then counted the number of bytes in that file, and generated an equally sized random binary data file. Finally, I added the bitmap header to the file, converted it to a PNG file, optimized it, and uploaded it here. The steps are as follows: $ gimp # create 600x600px plain white file and save as 16-bit "white.bmp"
$ls -l white.bmp | awk '{print$5}'
720138
$tail -f /run/rtl_entropy.fifo| dd of=random.img bs=1 count=720138 iflag=fullblock 720138+0 records in 720138+0 records out 720138 bytes (720 kB) copied, 24.8033 s, 29.0 kB/s$ dd if=white.bmp of=random.img bs=1 count=54 conv=notrunc
$gimp random.img # export to PNG file When viewing the output, there should be no obvious patterns in the output. As an example: For more practical use, here is a quick application for generating 80-bit entropy unambiguous passwords: $ for i in {1..10}; do
> strings /run/rtl_entropy.fifo | grep -o '[a-hjkmnp-z2-9.]' | head -n 16 | tr -d '\n'; echo
> done
8dfn42w6dagqnt4z
2vcsqu6sew.g6pp2
kv9nstj4gq39x5f.
wmpdpy7yz75xrhkh
.4ra2b38hmbf5jw5
7ngyk3c58k3eeq7c
8e4t8ts3ykhckdst
9g6yqqce.bxrrhpb
xwnw6mtk8njv76b2
xdmd89n68f.kcthp

Obviously, the practical uses here can be for Monte Carlo simulations, game theory, gambling, cryptography, and other practical uses where high quality randomness is needed. Unfortunately, I can seem to get rngd(8) to add the /run/rtl_entropy.fifo file as a hardware device. So, I can't feed the Linux CSPRNG with with the dongle, other than "dd if=/run/rtl_entropy.fifo of=/dev/random", which doesn't increase the entropy estimate, of course.

## Encrypting Combination Locks

This morning, my family and I went swimming at the community swimming center. Unfortunately, I couldn't find my key-based lock that I normally take. However, I did find my Master combination lock, but couldn't recall the combination. Fortunately, I knew how to find it. I took this lock with me to lock my personal items in the locker while swimming around in the pool.

While swimming, I started thinking about ways to better recall lock combinations in the future. The obvious choice is to encrypt it, so I could engrave the encrypted combination on the lock. However, it needs to be simple enough to do in my head should I temporarily forget it while swimming, and easy enough to recall if I haven't used the lock in a few years. Thankfully, this can be done easily enough with modulo addition and subtraction.

Before beginning, you need a 6-digit PIN that you won't easily forget. Tempting enough, dates can easily be in 6-digits, and something like a birthday or an anniversary are not hard to remember. Unfortunately, if someone knows you, and knows these dates, they can easily reverse the process to open the lock. So, as tempting as dates are, don't use them. Instead, you should probably use a 6-digit PIN, that only you would know, and always know. So, knowing this, let's see how this works.

You need to be familiar with modulus math, aka "clock math". The idea, is that after a certain maximum, the numbers reset back to 0. For example, 00:00 is midnight while 23:59 is the minute before. As soon as the hour is "24", then it resets back to 0, for a full 24-hour day. You could call telling time "mod 24 math". For combination locks, we're going to be using "mod 40 math", if the maximum number on your combination lock is "40", on "mod 60 math" if the max is "60", and so forth.

Suppose the combination to your lock is "03-23-36", and suppose your 6-digit PIN is "512133". Let's encrypt the combination with our PIN, by using "mod 40 subtraction". We'll use subtraction now, because most people have an easier time with addition than subtraction. When you are trying to rediscover your combination, you'll take your encrypted number, and do "mod 40 addition" to reverse it, and bring it back to the original combination lock numbers.

Here it is in action:

Encrypting the original combination

03 23 36    <- original combination
- 51 21 33    <- secret PIN
= --------
-48 02 03
= --------
32 02 03    <- encrypted after "mod 40"

Because the first number in our combination is "03", and we are subtracting off "51", we end up with "-48". As such, we need to add "40" until our target new number is in the range of [0, 40), or "0 <= n < 40". This gives us "32" as the result. The rest of the numbers fell within that range, so no adjusting was necessary. I can then engrave "32-02-03" on the bottom of the lock, so when I hold the lock up while in a locker, the text is readable. Okay, that's all fine and dandy, but what about reversing it? Taking the encrypted combination, and returning to the original combination? This is where "mod 40 addition" comes in. For example:

Decrypting the encrypted combination

32 02 03    <- encrypted combination
+ 51 21 33    <- secret PIN
= --------
83 23 36
= --------
03 23 36    <- original combination after "mod 40"

Notice that this time, the first number in our "mod 40 addition" is "83". So, we subtract of "40" until our original combination number is in the range of [0,40), or "0 <= n < 40", just like when doing "mod 40 subtraction" to create the new combination lock values. At worst case, you'll have to subtract a "40" only three times per number. On thing to watch out for, is that your encrypted combination numbers are far enough away from the original, that trying out the encrypted combination, won't accidentally open the lock, due to their proximity to the original numbers. If only one number is substantially off, that should be good enough to prevent an accidental opening. I want to come back to dates however, and why not to use them. Not only do they fall victim to a targeted attack, but they also have an exceptionally small key space. Assuming there are only 365 days per year, and assuming the attacker has a good idea of your age, plus or minus five years, that's a total of 3,650 total keys that must be tried following the common convention of "MM-DD-YY". It could be greatly reduced, if the attacker has a better handle on when you were born. If a 6-digit PIN is chosen instead, then the search space has 1,000,000 possible PINs. This is greater than the 64,000 possible maximum combination numbers a 40-digit Master lock could have, which puts the attacker on a brute force search for the original combination, if they aren't aware that Master combination locks can be broken in 8 tries or less.

## The Lagged Fibonacci Generator

Lately, I have been studying pseudorandom number generators (PRNGs, also called "deterministic random bit generators", or DRBGs). I've been developing cryptographically secure PRNGs (CSPRNGs), and you can see my progress on Github at https://github.com/atoponce/csprng. This project is for nothing more than for me to somewhat get a feeling for new languages, while also learning a thing or two about applied cryptograhpy. However, for the subject of this post, I want to address one PRNG that is not cryptographically secure- the Lagged Fibonacci Generator.

What drew me to this generator was thinking about a way to have a PRNG to do by hand. I started thinking about different ways to construct a PRNG mathematically. But, before creating an algorithm, I needed to identify all the points that make a good PRNG. A good PRNG should have:

• An easy implementation.
• High efficiency in calculating the pseudorandom values.
• Long (practically un-observable) periods for most, if not all initial seeds.
• A uniform distribution over the finite space.
• No correlation between successive values.

I put a great deal of thought into it, but couldn't come up with anything I was very proud of. I thought of using trigonometric functions, various logarithm functions, geometric and algebraic expressions, and even fancy equations using derivatives. The more I thought about it, the further away I drifted from something simple that could be done by hand with pencil and paper.

The best I came up with, which required using a scientific calculator, was forcing the sequence to grow (a monotonically increasing function), then forcing it into a finite field with a modulus. However, no matter what I threw at it, I always struggled with either dealing with "0" or "1". For example, taking the n-th exponent of either "0" or "1" will always return a "0" or "1". I realized quickly that multiplication might be a problem. For example, one thought I had was the following:

Si = Floor[(Si-1)3/2], mod M

This works out fine, until your output is a "0" or "1", then the generator sits on either of those numbers indefinitely. I realized that my function should probably just stick with addition, or I'm bound to get myself into trouble. I thought, and thought about it, then it hit me. It was my "Beautiful Mind" moment.

I thought of the Fibonacci sequence.

The Fibonacci sequence is monotonically increasing for two seeds S1 and S2, where 0 < S1 < S2. If you put an upper bound on the sequence via a modulus, you can limit it to a finite space, and I can have my PRNG. However, I also know that the distance between any two sequential digits in the Fibonacci sequence approaches the Golden Ratio Phi. I'm not sure how this would affect my simple PRNG, and if a correlation between successive digits could be identified, but I started scribbling down numbers on a text pad anyway.

Immediately, however, I found something interesting: If both seeds are even, then the whole sequence of numbers would be even. For example, take the following Fibonacci PRNG:

S1 = 6, S2 = 8, mod 10
6 8 4 2 6 8 4 2 6 8 4 2 ...

There are two problems happening here- first, the period of the PRNG is 4 digits- 6, 8, 4, & 2. Second, because even numbers were chosen for the seeds, even numbers are the only possibility for the PRNG. So, either one of the seeds or the modulus must be odd, or the PRNG algorithm needs to be modified.

At this point, I threw my hands up in the air, and said "screw it". I decided to see what history had discovered with simple PRNGs. Turns out, I wasn't far off. A Fibonacci sequence PRNG exists called the Lagged Fibonacci Generator. Here is how it works:

Sn = Sn-j ⊙ Sn-k mod M, 0 < j < k

Where "⊙" is any binary function, such as addition, subtraction, multiplication, or even the bitwise exclusive-or.

First off, it doesn't address the "all evens" problem with my naive generator. If addition is used to calculate the values, then at least one number in the seed must be odd. If multiplication is used, then at least k-elements must be odd. However, what is interesting about this generator, is that rather than picking the first and second elements of the list to calculate the random value (Si-1 and Si-2), any j-th and k-th items in the list can be used (Si-j and Si-k). However, you must have at least k-elements in the list as your seed before beginning the algorithm.

To simplify things, lets pick "j=3" and "k=7" mod 10 addition. I need at least seven elements in the list, and at least one of them must be odd. I've always like the phone number "867-5309", so let's use that as our seed. Thus, the first 10 steps of our generator would look like this:

j=3, k=7, mod 10 addition

[j]       [k]
1. 8 6 [7] 5 3 0 [9] => 7+9 = 6 mod 10
2. 6 7 [5] 3 0 9 [6] => 5+6 = 1 mod 10
3. 7 5 [3] 0 9 6 [1] => 3+1 = 4 mod 10
4. 5 3 [0] 9 6 1 [4] => 0+4 = 4 mod 10
5. 3 0 [9] 6 1 4 [4] => 9+4 = 3 mod 10
6. 0 9 [6] 1 4 4 [3] => 6+3 = 9 mod 10
7. 9 6 [1] 4 4 3 [9] => 1+9 = 0 mod 10
8. 6 1 [4] 4 3 9 [0] => 4+0 = 4 mod 10
9. 1 4 [4] 3 9 0 [4] => 4+4 = 8 mod 10
10. 4 4 [3] 9 0 4 [8] => 3+8 = 1 mod 10

Generated: 6 1 4 4 3 9 0 4 8 1

The following Python code should verify our results:

 123456789101112 j = 3 k = 7 s = [8, 6, 7, 5, 3, 0, 9] for n in xrange(10):     for i in xrange(len(s)):         if i is 0:             out = (s[j-1] + s[k-1]) % 10 # the pseudorandom output         elif 0 < i < 6:             s[i] = s[i+1] # shift the array         else:             s[i] = out             print s[i], # print the result

Running it verifies our results:

$python lagged.py 6 1 4 4 3 9 0 4 8 1 It's a "lagged" generator, because "j" and "k" lag behind the generated pseudorandom value. Also, this is called a "two-tap" generator, in that you are using 2 values in the sequence to generate the pseudorandom number. However, a two-tap generator has some problems with randomness tests, such as the Birthday Spacings. Apparently, creating a "three-tap" generator addresses this problem. Such a generator would look like: Sn = Sn-j ⊙ Sn-k ⊙ Sn-l mod M, 0 < j < k < l Even though this generator isn't cryptographically secure (hint: it's linear), it meets the above requirements for a good PRNG, provided the "taps" are chosen carefully (the lags are exponents of a primitive polynomial), and the modulus is our traditional "power-of-2" (2M, such as 232 or 264). Supposing we are using a two-tap LFG, it would have a maximum period of: (2k-1)*k if exclusive-or is used (2k-1)*2M-1 if addition or subtraction is used (2k-1)*2M-3 if multiplication is used (1/4 of period of the additive case) For a good LFG, it is found that a three-tap generator should be used, as a 3-element spacing correlation can be found in two-tap generators, and that initial taps should be very high for a large modulus. Further, the full mathematical theory hasn't been worked out on Fibonacci generators, so the quality of the generators rests mostly on the statistics of the generated output, and randomness tests. However, this is simple enough to do by hand, if nothing else than to impress your friends. ## Financially Supporting Open Crypto In April 2014, Heartbleed shook the Internet. OpenSSL had introduced a feature called "TLS Heartbeats" Heartbeats allow for a client-encrypted session to remain open between the client and the server, without the need to renegotiate a new connection. In theory, the feature is sound. Heartbeats should minimize load on busy servers, and improve responsiveness on the client. However, due to a simple oversight in the code, buffers could be over-read, allowing the client to request much more data from the server's memory than needed. As a result, usernames and passwords cached in the server's memory could be leaked to the client. This was a nasty bug, and it underscored how under-staffed and under-funded the OpenSSL development team is. OpenSSL is the de facto standard in securing data in motion for the Internet. It protects your web connections when visiting your bank's website, and it protects your email communication between your email client and the upstream mail server. Ars Technica started off an article about tech giants finally agreeing to fund the OpenSSL development. Quote: The open source cryptographic software library secures hundreds of thousands of Web servers and many products sold by multi-billion-dollar companies, but it operates on a shoestring budget. OpenSSL Software Foundation President Steve Marquess wrote in a blog post last week that OpenSSL typically receives about$2,000 in donations a year and has just one employee who works full time on the open source code.

If that isn't bad enough, Werner Koch, the sole developer and maintainer of the encryption software "GnuPG" is in much the same position as Steve Marquess. ProPublica put up a post about the very sobering financial situation of GnuPG. Quote:

The man who built the free email encryption software used by whistleblower Edward Snowden, as well as hundreds of thousands of journalists, dissidents and security-minded people around the world, is running out of money to keep his project alive.

Werner Koch wrote the software, known as Gnu Privacy Guard, in 1997, and since then has been almost single-handedly keeping it alive with patches and updates from his home in Erkrath, Germany. Now 53, he is running out of money and patience with being underfunded.

To understand just how critical this piece of software is to the Internet and the community at large, OpenPGP (the specification upon which GnuPG is built) is used by software developers around the world to prove the integrity of their software, when downloading it from their website. It's used by operating system vendors, such as Microsoft, Apple, Google, and GNU/Linux to provide package integrity when installing "apps" on your computer or mobile device. People and corporations have used it internally for data at rest as well, such as encrypting backups before sending them offsite.

Thankfully, after ProPublica published their article, Werner Koch, father and husband, got the donation funding he needed to continue focusing on it full time. Thanks to Facebook and Stripe, he has $100,000 of annual sponsored donations to help keep the development of GnuPG pressing forward. Why is it that the two most fundamental cryptographic tools in our community are so under developed, under funded, and under staffed? I can understand that cryptography is hard. There is a reason why people get doctorate degrees in mathematics and computer science to understand this stuff. But with such critical pieces of infrastructure protection, you would think it would be getting much more attention than it is. A good rule of thumb for cryptography, is if you want to protect your data in transit, use OpenSSL; if you want to protect your data at rest, use GnuPG. Let's hope that these two projects get the attention and funding they need to continue well into the future for years to come. If you want to help donate to these two projects, you can donate to GnuPG here and to OpenSSL here. Alternatively, there is a Flattr donation page for GnuPG where you can setup recurring donations here. ## Reasonable SSH Security For OpenSSH 6.0 Or Later As many of you have probably seen, Stribik András wrote a post titled Secure Secure Shell. It's made the wide rounds across the Internet, and has seen a good, positive discussion about OpenSSH security. It's got people thinking about their personal SSH keys, as well as the differences between ECC and RSA, why the /etc/ssh/moduli file matters, and other things. Because of that post, many people who use SSH are increasing their security when they get online. However, the post does one disservice- it requires OpenSSH 6.5 or later. While this is good, and people should be running the latest stable release, there are many, many older versions of OpenSSH out there, that are still supported by the distro, such as Debian GNU/Linux 7.8, which ships OpenSSH 6.0. Most people will be using the release that ships with their distro. As a side note, CentOS 5 ships OpenSSH 4.3, and CentOS 6 ships OpenSSH 5.3. Because these are very old releases, and CentOS is still providing support for them, you will need to check the man pages for OpenSSH, and see how your client and server configurations need to be adjusted. It won't be covered here. So, with that in mind, let's look at OpenSSH 6.0, and see what it supports. OpenSSH 6.0 Ciphers The following is the default order for symmetric encryption ciphers: 1. aes128-ctr 2. aes192-ctr 3. aes256-ctr 4. arcfour256 5. arcfour128 6. aes128-cbc 7. 3des-cbc 8. blowfish-cbc 9. cast128-cbc 10. aes192-cbc 11. aes256-cbc 12. arcfour CTR mode should be preferred over CBC mode, whenever possible. It can be executed in parallel, and it seems to be the "safer" choice over CBC although it's security margin over CBC is probably minimal. The internal mechanisms are more simplistic, which is why modes like EAX and GCM use CTR internally. With that said, CBC mode is not "unsafe", so there is no strong security argument to avoid it. However, modern and older OpenSSH implementations support CTR mode, so there really is no need for CBC. The "arcfour" protocols are "alleged RC4", but adhere to the RC4 RFC. RC4 has been showing weaknesses lately. Cryptographers have been advising to move off of it, PCI vendors will fail scans with SSL implementations that support RC4, and OpenBSD 5.5 switched to a modified ChaCha20 for its internal CSPRNG. So, it's probably a good idea to move away from the arcfour ciphers, even if it may not be practically broken yet. However, arcfour is really the only high performance cipher in the OpenSSH 6.0 suite, and is very handy when trying to transfer many gigabytes of data over the network, as AES will pin the CPU before flooding the pipe (unless of course you have hardware AES on board). So, I would recommend the arc4 ciphers as a last resort, and only enable them on private networks, where you need the throughput. The cast128 cipher was an AES candidate, and is a Canadian standard. To my knowledge, it does not have any near practical security attacks. However, because only CBC mode is supported with CAST, and not CTR mode, and we're disabling CBC mode, it is not included in our final list. 3DES was designed to address the short 56-bit key sizes in DES, which was replaced later by AES. 3DES cascades DES three times, with three distinct 56-bit keys. 3DES also does not have any near practical security attacks, and it is believed to be secure. However, DES was designed with hardware in mind, and is slow, slow, slow in software. 3DES three times as much. It's horribly inefficient. As such, I would recommend disabling 3DES. Blowfish was designed by Bruce Schneier as a replacement for DES. While Blowfish might still have a considerable security margin, Blowfish suffers from attacks from weak keys. As such, Blowfish implementations must be careful when selecting keys. Blowfish can be efficient in both hardware and software, but it's usually less efficient than AES. Further, Bruce himself recommends that people stop using Blowfish and move to its successor Twofish, or even Threefish. As such, because both stronger and more efficient algorithms exist, I would recommend disabling Blowfish. It really isn't offering anything to OpenSSH clients. So, in my opinion, I would sort my OpenSSH 6.0 ciphers like so: 1. aes256-ctr 2. aes192-ctr 3. aes128-ctr 4. arcfour256 5. arcfour128 6. arcfour OpenSSH 6.0 Key Exchange The following is the default order for key exchange algorithms: 1. ecdh-sha2-nistp256 2. ecdh-sha2-nistp384 3. ecdh-sha2-nistp521 4. diffie-hellman-group-exchange-sha256 5. diffie-hellman-group-exchange-sha1 6. diffie-hellman-group14-sha1 7. diffie-hellman-group1-sha1 The NIST curves are considered to be insecure. Not because it's some government agency tied with the NSA, but because the curves are not ECDLP rigid, and suffer from a lack of constant-time single-coordinate single-scalar multiplication, they aren't complete, and are distinguishable from uniform random strings. If you want to blame the NSA for rubber-stamping and backdooring the NIST ECC curves, fine. I'll stick with the crypto. And, although the security margin gap is closing on SHA-1, some commercial SSH providers, such as Github may still require it for your SSH client. So, in your client config, I would put the preference on SHA-256 first, followed by SHA-1. On your own personal servers, you can disable the SHA-1 support completely. Thus, I would recommend the following key exchange order: 1. diffie-hellman-group-exchange-sha256 2. diffie-hellman-group-exchange-sha1 3. diffie-hellman-group14-sha1 4. diffie-hellman-group1-sha1 OpenSSH 6.0 Message Authentication Codes The following is the default order for message authentication codes: 1. hmac-md5 2. hmac-sha1 3. umac-64@openssh.com 4. hmac-ripemd160 5. hmac-sha1-96 6. hmac-md5-96 7. hmac-sha2-256 8. hmac-sha256-96 9. hmac-sha2-512 10. hmac-sha2-512-96 Things get interesting here, because with HMAC algorithms, successful attacks require breaking the preimage resistance on the cryptographic hash. This requires a complexity of 2^n, where "n" is the output digest size in bits. MD5 is 128-bits, and SHA-1 is 160-bits. All currently known attacks on MD5 and SHA-1 are collision attacks, and not preimage attacks. Collision attacks require a complexity of only 2^(n/2). Thus, for MD5, collision attacks require a complexity of only 64-bits at worst, and SHA-1 requires 80-bits. However, as we know now, MD5 collision resistance is fully broken in practical time with practical hardware. SHA-1 still remains secure, although its collision resistance has been weakened to 61-65-bits. This is almost practical. Regardless, the HMAC-MD5 and HMAC-SHA1 remain secure, with wide security margins, due to their preimage resistance. The only concern, however, is that in order to succesfully break the preimage resistance of a cryptographic hash function, it requires first breaking its collision resistance. Because MD5 is broken in this regard, and SHA-1 is almost broken, it is advised to move away from any protocol that relies on MD5 or SHA-1. As such, even though HMAC-MD5 and HMAC-SHA1 remain very secure today, it would be best to disable their support. Interestingly enough, even though RIPEMD-160 has the same digest output space as SHA-1, it has no known collision weaknesses, and remains secure today, almost 20 years since its introduction. Due to the almost practical collision attacks on SHA-1 with a a complexity of 61-65 bits, UMAC-64 probably does not have a wide enough security margin. As such, it should probably be disabled. I would recommend the following order for your MACs: 1. hmac-sha2-512 2. hmac-sha2-256 3. hmac-ripemd160 OpenSSH 6.0 Configuration Okay. Now that we've everything ironed out in hardening our OpenSSH 6.0 connections, let's see how this would look in the client and on the server. For both the client config and the server config, it should support algorithms for both OpenSSH 6.0 and 6.7. For an OpenSSH 6.0 client, I would recommend this config: # OpenSSH 6.0 client config Host * Ciphers aes256-ctr,aes192-ctr,aes128-ctr,arcfour256,arcfour128,arcfour KexAlgorithms diffie-hellman-group-exchange-sha256,diffie-hellman-group-exchange-sha1,diffie-hellman-group14-sha1,diffie-hellman-group1-sha1 MACs hmac-sha2-512,hmac-sha2-256,hmac-ripemd160  For an OpenSSH 6.0 server, I would recommend this config: # OpenSSH 6.0 server config Ciphers aes256-ctr,aes192-ctr,aes128-ctr,arcfour256,arcfour128,arcfour KexAlgorithms diffie-hellman-group-exchange-sha256 MACs hmac-sha2-512,hmac-sha2-256,hmac-ripemd160  Going back now to Stribik András' post, here is what your configurations would look like for OpenSSH 6.7: For an OpenSSH 6.7 client, I would recommend this config. Further, ChaCha20-Poly1305 is a high performance cipher, similar to RC4. So we should prefer it as our first cipher, with AES following, and finally disabling RC4: # OpenSSH 6.7 client config Host * Ciphers chacha20-poly1305@openssh.com,aes256-gcm@openssh.com,aes128-gcm@openssh.com,aes256-ctr,aes192-ctr,aes128-ctr KexAlgorithms curve25519-sha256@libssh.org,diffie-hellman-group-exchange-sha256,diffie-hellman-group-exchange-sha1,diffie-hellman-group14-sha1 MACs hmac-sha2-512-etm@openssh.com,hmac-sha2-256-etm@openssh.com,hmac-ripemd160-etm@openssh.com,umac-128-etm@openssh.com,hmac-sha2-512,hmac-sha2-256,hmac-ripemd160,umac-128@openssh.com  For an OpenSSH 6.7 server, I would recommend this config (also disabling SHA-1 from the key exchanges): # OpenSSH 6.7 server config Ciphers chacha20-poly1305@openssh.com,aes256-gcm@openssh.com,aes128-gcm@openssh.com,aes256-ctr,aes192-ctr,aes128-ctr KexAlgorithms curve25519-sha256@libssh.org,diffie-hellman-group-exchange-sha256 MACs hmac-sha2-512-etm@openssh.com,hmac-sha2-256-etm@openssh.com,hmac-ripemd160-etm@openssh.com,umac-128-etm@openssh.com,hmac-sha2-512,hmac-sha2-256,hmac-ripemd160,umac-128@openssh.com  Conclusion It's important that you pay attention to the versions of the clients and servers that you are using, so you can accurately set your configuration. In this case, we looked at what would be necessary to support OpenSSH versions 6.0 and 6.7. There may be slight differences in versions between those two, and you'll need to make the necessary adjustments. ## Verifying Keybase Identities When using Keybase, occasionally, people will track your identity. This has cryptographic value. Your identity on Keybase is based on what you do online and how long you have done it. As people track you, they cryptographically sign your Keybase identity. This creates a snapshot in time that states you've taken the precautions to verify the identity, by checking the digital signature of each of their online proofs. This snapshot is frozen in time, and as more and more people track your identity, the stronger the statement of the validity of that identity. In other words, Keybase compliments the PGP Web of Trust, without actually replacing key signing parties, or actually signing PGP keys. In this post, I want to discuss what it takes to verify signatures of Keybase identity proofs, so you can verify that Keybase isn't doing anything sneaky the data. In this post, I am going to verify the identity proofs of a friend of mine, Joshua Galvez as an example of how to verify each identity proof out-of-band (not using the Keybase client software). First, all identity proofs are stored in JSON, which is a standardized format. The JSON object is cleanly formatted for easy readability, so you can examine what has been signed, and exactly what you are verifying. Nothing should be hidden up Keybase's sleeves. To start, I am going to navigate to Josh's Keybase identity page. I see that he has proved he owns a Twitter account, a Github account, a reddit account, and a personal website, all with his personal OpenPGP key. To verify the proofs, I need to get a physical copy of the statement. Again, I am going to do this all out-of-band, away from the Keybase client software. As such, I'll copy and paste each statement proof into a text editor, and save it to disk, as well as each PGP signature. I'll do this with his Twitter account as an example. Because of the brevity of Twitter, a full JSON object with a PGP signature can't be sent. So, Keybase keeps this proof on their server, with a link in the tweet pointing to the proof. So, we'll need to get it there. The link in his tweet points to https://keybase.io/zevlag/sigs/0Pl859RFLHZuEi7ozQyrbT1cphZCxYQMuoyM. There is a "Show the proof" link on the page, which gives me all the necessary data for verifying his identity. All I need is his JSON object and his PGP signature. I need to combine them in a single file, and save it to disk. As such, my file will look like this: { "body": { "client": { "name": "keybase.io node.js client", "version": "0.7.3" }, "key": { "fingerprint": "12c5e8619f36b0bb86b5be9aea1f03e20cf2fdbd", "host": "keybase.io", "key_id": "EA1F03E20CF2FDBD", "uid": "2b26e905f5b23528d91662374e840d00", "username": "zevlag" }, "service": { "name": "twitter", "username": "zevlag" }, "type": "web_service_binding", "version": 1 }, "ctime": 1416507777, "expire_in": 157680000, "prev": null, "seqno": 1, "tag": "signature" } -----BEGIN PGP MESSAGE----- Version: GnuPG/MacGPG2 v2.0.22 (Darwin) Comment: GPGTools - https://gpgtools.org owGbwMvMwMVYnXN9yvZMuQbG0wdeJjGE5BnOrlZKyk+pVLKqVkrOyUzNKwGx8hJz U5WslLJTK5MSi1P1MvMV8vJTUvWyihWganSUylKLijPz84CqDPTM9YyVanVAykGa 0zLz0lOLCooyQWYpGRolm6ZamBlaphmbJRkkJVmYJZkmpVompiYaphkYpxoZJKcZ paUkpQCNzMgvLkGxVQlsZnxmClDU1dHQzcDY1cjA2c3IzcXJBShXCpYwSjIyS7U0 ME0zTTIyNjWySLE0NDMzMjY3SbUwMUgxMAApLE4tgnqpKrUsJzEd5FqgWFlmciqS d0vKM0tKUotwaSipLAAJlKcmxUP1xidl5qUAfYscHIZAlcklmSC9hiaGZqYG5kCg o5RaUZBZlBqfCVJham5mYQAEOkoFRallSlZ5pTk5IPcU5uUDZYEWAe2zUirOTM9L LCktSlWq7WSSYWFg5GJgY2UCxRgDF6cALB6Z5wowNEXaavBftOToU3lx8YCcUIzU Y5FN3LxXjrbNOum2oiQyrMMszfWM5Kz+D3O5ZZJabOUO2v99UPBMRmqZy6ZLdh0n 1t92OPT+7ILwL2+Y+rfoLDJxufXh7ykfZos1L66fVhc+e1HOw9rEuChW+eBJkbCE 3y42k3yXNJ5sSpDc9Ujhcewsw3nuM86G/8tbzGUo9OSERcif0wou4Qtnnj2cs6Nz nk3qLUvHWRfX753v+mPlNMm2hXZbbjzIF3XaK93JZj/5im6QyA3fjy6CbKGtqi1X Am+d+ljtGD0h5MPRDLPG3Xtmiqps3Hfo7fGvUUwF6gnzHm7xdD42banr8wMHfZ6q a26o3nPras+GHwERx898vvTq14lfxgKc7wXOXEl30bN7KTa/TERMXK2nafZT9qOb 7sSvSnjLsj9jw1nVGaKcvtd8855Xr6/eySwtsjPOapVT85JPisu3l/ten/XqxPa8 mRxWAvrLdl9NXNS46zYPW61MwJQPYQfYN3DxTX3ucMO5Qd/EftWeF0depDlsme7S pTmr0Eey8sqkPo9f63/yfFgg9GPP3PY/22KP7+2Ifm7/N6FZ95xh2bXXM9dPvJ79 gCVVbH5m4OKOTI3aD9bWnVekglJLpwboJb6dfPC4V0vmpgWxb0w49JfE3j8fHHNa jTnvgo3T+vzZ4mcOG06+/2CpebSKtmlTyIyYPzYbDXWczSPVPN28Yhr7rwIA =V5X5 -----END PGP MESSAGE----- I'll save this to disk as /tmp/zevlag-twitter.txt Now, I just need Josh's public PGP key imported from a key server. I can, and should use Keybase here. Instead of using the MIT PGP key server, and running the risk of getting the wrong key, I can be reasonably confident I will get the correct key from Keybase. The raw public key can be accessed by appending "key.asc" at the end of their identity URL. So, in this case https://keybase.io/zevlag/key.asc. So, I'll grab it via the shell: $ wget -O - https://keybase.io/zevlag/key.asc 2> /dev/random | gpg --import -

Now that I have Josh's public key imported into my GPG public key ring, I am read to verify Josh's Twitter proof of identity:

$gpg --verify /tmp/zevlag-twitter.txt gpg: Signature made Thu 20 Nov 2014 11:23:23 AM MST using RSA key ID B7691E80 gpg: Good signature from "Joshua Galvez <josh@zevlag.com>" gpg: aka "Joshua Galvez (Work - Emery Telcom) <jgalvez@emerytelcom.com>" gpg: aka "keybase.io/zevlag <zevlag@keybase.io>" gpg: WARNING: This key is not certified with a trusted signature! gpg: There is no indication that the signature belongs to the owner. Primary key fingerprint: 12C5 E861 9F36 B0BB 86B5 BE9A EA1F 03E2 0CF2 FDBD Subkey fingerprint: DC35 E3CF 1179 41A9 7D72 BC9A 7B6C D794 B769 1E80 At this point, I can confirm that the owner of the private key for 0xEA1F03E20CF2FDBD cryptographically signed a JSON object for Twitter. Further, that individual has access to the Twitter account, so the signature can be posted. After verifying the other accounts, I can be reasonably confident that the individual is who they claim- Josh Galvez. Otherwise, an attacker has successfully compromised all of Josh Galvez's online accounts, as well as his OpenPGP key (or forged a new one), and either compromised his Keybase account, or created one masquerading as him. The former seems more likely than the latter. Further, because I have previously met with and engaged online with Josh, I have no doubt that this is indeed Josh Galvez, and 0xEA1F03E20CF2FDBD is indeed his public key. So, I can now track Josh through Keybase, which means me cryptographically signing his Keybase identity, and creating a snapshot in time that says "I am reasonably sure this is Josh Galvez, these accounts are part of his online presence, and 0xEA1F03E20CF2FDBD is his OpenPGP key. Staying out of band from the Keybase client software, I can do this entirely with curl(1) and gpg(1). Navigating to his Keybase identity, I'll click the "Track zevlag" button. A pop-up displays with the following options: • in the browser • command line with keybase • command line with [bash + GPG + cURL] I have not integrated an encrypted copy of my private key with Keybase, so tracking Josh in the browser is unavailable to me. Further, I wish to do this out-of-band from Keybase anyway, so I'll select "command line with [bash + GPG + cURL]" and click "Continue". This displays that I need to copy and paste the following content into my shell: echo '{"body": (... large JSON object snipped ....) }' | \ gpg -u 'e0413539273a6534a3e1925922eee0488086060f' -a --sign | \ perl -e '$_ = join("", <>); s/([^\w\.@ -])/sprintf("%%%2.2x",ord($1))/eg; s/ /+/g; print("sig=",$_)' | \
curl -d @- \
-d type=track \
-d session=lgHZIDg3ZWNjY2NiNTRiMTBiNThjOTQ2NDJhODA3MzM2NjAwzlSh4WnOAeEzgNkgZjZmNWVmZDg4YzcwZDI2NDNlZGY2ZWYyYTc3M2IyMDLEIM0QqHGrtfga4a%2Bnz7soXFHqFbbiio7PaVGjh7DfyyPG \
-d plain_out=1 \
-d uid=2b26e905f5b23528d91662374e840d00 https://keybase.io/_/api/1.0/follow.json

After entering that into my shell, and hitting enter, I am presented with typing in my passphrase for my private key, which in turn signs the object, and uses the Keybase API to post the result. I can then reload my profile, and see that I am now tracking Josh with Keybase. This means that at this point in time, I have made a cryptographic statement regarding the key ownership and identity of Joshua Galvez. Of course, I can revoke that statement at any time, if for any reason I believe his account has become compromised, he himself has become untrustworthy, or for other reasons.

## Keybase and The PGP Web of Trust

Recently, I have been playing with my Keybase account, and I thought I would weigh in on my thoughts about it compared to the PGP Web of Trust (WoT).

The PGP WoT tries to solve the following two problems directly:

1. You have the correct key of the person to whom you wish to communicate.
2. You have verified that the owner of that key is who they claim to be.

These two problems are solved through key signing parties. Two or more people will meet up, exchange key fingerprints, then verify personal identity, usually through government issued identification. Unfortunately, the PGP WoT is complex, and in practice, rarely, if ever used. The idea behind using the PGP WoT is this:

• I have verified Adam's identity and confirmed I have his correct key.
• I cryptographically signed his key as a statement of this verification.
• Adam cryptographically signed Bruce's key, issuing a similar statement.
• I haven't met Bruce, but I have met Adam, and trust him.
• Through Adam, I can make a statement about Bruce's claim to identity.

In practice, if I wished to communicate securely with Bruce, I would see if Bruce's key has signatures of individuals that I have cryptographically signed. If so, I can make a weak statement about his identity, and the ownership of his key through that signature. From that standpoint, I can then determine if I wish to communicate securely with Bruce, or not.

Since using GnuPG these past 10 years, I have probably really used the PGP WoT only 2-3 times. Other than that, it makes for a sweet-looking directed graph.

Keybase is not a PGP WoT replacement. IE, it's not here to replace key signing parties, and it's not a tool for signing other's keys. However, Keybase does make strong statements regarding key ownership and identity. In fact, Keybase has given up on the PGP Wot entirely. Rather than validating government issued identification cards in person, Keybase solves identity through online social proofs. This is handled by what you have accomplished online and how long you have been using the account.

Looking first and accomplishing online tasks. When a user signs up for an account at Keybase, they need to prove identities that they own on the web. This is done by inserting some text at the online account, then cryptographically signing it with your private PGP key, and storing the signature at Keybase. This establishes a relationship between the owner of the PGP key and the online account. The more online accounts that the user can establish, the stronger the proof of identity for that individual.

Currently, accounts can be:

• Reddit
• Hacker News
• Coinbase
• Github
• Websites

For each of these accounts, I can pull down the notice, and verify the signature. Thus, each online account becomes coupled with the owner's PGP key. But, it's important to understand that this is making a statement of online activity. IE- "This is my Twitter account @AaronToponce, and I am Aaron Toponce."

Once the accounts have been proved, you can then make statements about other identities through "tracking". Tracking on Keybase is similar to "following" on other social sites, but it's actually cryptographically useful. Each account has a database object of their online identities (all cryptographically signed remember), among other data, including who they are tracking, and who is tracking them.

When you track someone, you cryptographically sign their identity with your personal PGP key. The previous signature is part of that identity, as well as the current signature. Each time someone is tracked, their identity gets cryptographically updated, and anyone can see when those signatures took place. Think of tracking like cryptographic snapshots, or digital photographs.

Tracking is useful for people whom you wish to communicate, are interested in "following" them online. By looking at the previous snapshots, you can get a sense of the age of that account. The older the account, and the more people tracking the account, the stronger the statement of identity, and that the account has not been compromised. Should the account get compromised at any time, people can revoke their tracking snapshot, thus removing the statement of identity.

Will Keybase improve the overall PGP WoT? I hope so. Currently, the accounts that you can make verifiable proofs with are limited, and you'll notice the Big Players like Google, Facebook, and Pinterest are missing. Currently Keybase is in limited invite-only alpha testing, so it makes sense why those accounts are have not been brought into the system yet. However, Keybase will remain only a "geek it up" thing until those services are included in identity proofs. So, if Keybase wants to improve things with PGP in general, it must get those accounts on board, or it won't make a ripple in the world at large.

Oh, and the Keybase client is Free Software.

## SHA512crypt Versus Bcrypt

On the Internet, mostly in crypto circles, you'll see something like the following in a comment, forum post, on a mailing list, other otherwise:

Do not use fast hashes to store passwords on disk. Use bcrypt.

In most cases, however, the understanding of why to use bcrypt isn't entirely clear. You'll hear the standard answer "It's slow", without a real understand as to "how slow?" nor as to "why is it slow?". In this post, I want to explain why bcrypt is slow, some misconceptions about using fast hashes, and where the real strength of bcrypt lies (hint- it's not speed). Finally, I'll close with an alternative that many are starting to talk about as a possible replacement to bcrypt.

First, when people are talking about using bcrypt for password hashing, they are referring to the bcrypt cryptographic key derivation function, designed by Niels Provos and David Mazières. Bcrypt is designed to be intentionally slow and expensive. It was designed specifically with password storage in mind. The motivation is clear- if a password database of any kind is leaked to the Internet, it should be cost prohibitive for password crackers to make any sort of progress recovering the unknown passwords from the known hashes.

bcrypt algorithm
How does bcrypt work though? What is the algorithm? According to the paper, the core bcrypt function in pseudocode is as follows:

bcrypt(cost, salt, input)
state = EksBlowfishSetup(cost, salt, input)
ctext = "OrpheanBeholderScryDoubt" //three 64-bit blocks
repeat (64)
ctext = EncryptECB(state, ctext) //encrypt using standard Blowfish in ECB mode
return Concatenate(cost, salt, ctext)


The first function, "EksBlowfishSetup(cost, salt, input)" in the algorithm is defined as follows:

EksBlowfishSetup(cost, salt, key)
state = InitState()
state = ExpandKey(state, salt, key)
repeat (2^cost) // exponential cost by powers of 2
state = ExpandKey(state, 0, key)
state = ExpandKey(state, 0, salt)
return state


In the "EksBlowfishSetup", you'll notice the "repeat" step uses a binary exponential parameter. As the cost is increased, the time it will take to finish the algorithm will take exponentially longer. Bcrypt was designed with this cost parameter to adjust for Moore's law. As computing strength continues to improve, bcrypt should be flexible in its design to adjust for those advancements. This is why the cost parameter is baked into bcrypt, and why people call it "slow".

Finally, you'll notice the "ExpandKey(state, salt, key)" function in the algorithm. It is defined as follows:

ExpandKey(state, salt, key)
for(n = 1..18)
P_n  key[32(n-1)..32n-1] XOR P_n //treat the key as cyclic
ctext = Encrypt(salt[0..63])
P_1 = ctext[0..31]
P_2 = ctext[32..63]
for(n = 2..9)
ctext = Encrypt(ctext XOR salt[64(n-1)..64n-1]) //encrypt using the current key schedule and treat the salt as cyclic
P_2n-1) = ctext[0..31]
P_2n = ctext[32..63]
for(i = 1..4)
for(n = 0..127)
ctext = Encrypt(ctext XOR salt[64(n-1)..64n-1]) //as above
S_i[2n] = ctext[0..31]
S_i[2n+1] = ctext[32..63]
return state


Because bcrypt was based on Blowfish, the "ExpandKey(state, 0, key)" function used in the "EksBlowfishSetup" function is the same as regular Blowfish key schedule since all XORs with the all-zero salt value are ineffectual. The bcrypt "ExpandKey(state, 0, salt)" function is similar, but uses the salt as a 128-bit key.

Also, to clarify, a 128-bit salt is also baked into the algorithm, as you can see. This is to prevent the building of lookup tables for bcrypt, such as rainbow tables. Salts do not slow down crackers, and it's assumed that salts will be leaked with the database. All salts provide is the protection against using a hash lookup table to find the originating plaintext. Because salts are baked into bcrypt, bcrypt lookup tables will never exist. This forces password crackers to brute force the hash.

There are a few key security elements related to passwords that you must understand. They are the following:

1. The unpredictability measurement, aka "entropy", of the password provided by the user.
2. The speed at which brute forcing passwords can commence.
3. The cryptographic strength of the function itself.

I ordered these for a specific reason- the most likely "weak link" in the chain of password security is password the user provides. History of leaked password databases have shown us that. If users understood real strength behind passwords, they would understand the basic concepts of entropy, even if they weren't familiar with the term itself. If entropy levels were high in all user's passwords, no matter what, then the success of recovering passwords from hashes via brute force would be ineffective. But, 70-80%, and better, of password databases are recovered, because of this simple concept not getting applied. The speed at which password crackers brute forced their way through the hashes in the database would no longer matter, because no amount of practical computing power would be able to work fast enough within the death of the Universe, to recover the user's password.

Key stretching
So, we need to compensate for users picking bad passwords, and bcrypt makes a great leap in this regard. Because of the cost parameter which is part of the algorithm, we can adjust the cost to make password hashing intentionally slow. And, as computing power increases, the cost parameter can continue to be adjusted to compensate. This is what most people understand, when they claim that "bcrypt is slow".

The argument is that cryptographic hashes are designed to be fast, fast, fast. And they're right. Cryptographic hash functions are designed to provide data integrity regardless of the size of the input. If I have a 4.7 GB CD image, I should be able to calculate its digest in reasonable time, so when I transfer the image to another computer, I can recalculate the digest, and compare that the two digests match, in reasonable time.

This would seem like a Bad Thing for password storage, because passwords are short (much shorter than 4.7 GB at least), so password crackers would be able to guess millions or billions of passwords per second using a fast cryptographic hash. You're right, for the most part. Ars Technica ran a story on password cracking with a 25-GPU cluster. It achieves a speed of 350 billion NTLM passwords per second, which means every conceivable Windows NTLM password can be recovered in less than 6 hours using this behemoth. It can work MD5 at 180 billion per second, or 63 billion per second with SHA-1.

At these speeds, the argument against using fast cryptographic hashes to store passwords is sounding pretty strong. Except, that Ars Technica article, and most bcrypt fanboys seem to overlook one thing- key stretching. Just because cryptographic hashes are fast, doesn't mean we can't intentionally slow them down. This is key stretching.

Key stretching is the idea that you reuse internal state for calculating a new key. For cryptographic hash functions, this is "iterations" or "rotations". The idea is taking the cryptographic digest of the input, and using this digest as the input for another hashing round. In pseudocode, you could think of it like this:

salt = random() // programmatically determine a salt randomly
key = ''
cost = 5000

for ROUND in 1 to cost: do
key = digest
done


If our 25-GPU cluster could work through 50 billion SHA-512 cryptographic hashes per second, by forcing 5,000 SHA-512 calculations before getting to the desired hash, our 25-GPU cluster can now only work through 10 million SHA-512 hashes per second. As the iterative count is increased, the time it takes to calculate the resulting digest increases. As such, we have created a "sha512crypt" that has a similar cost parameter as bcrypt. Now the question remains- does it hold up?

Practical examples
To see if this "key stretching" idea holds up, I wrote two Python scripts- one using SHA-512, and the other using bcrypt. In both cases, I increase the cost parameter from a reasonable starting point, and increased it well beyond a reasonable expectation.

Here is my Python code for "test-sha512.py":

 123456789 #!/usr/bin/python import hashlib password = b'password' cost = 5000 key = '' m = hashlib.sha512() for i in xrange(cost):     m.update(key+password)     key = m.digest()

And here is my Python code for "test-bcrypt.py":

In both cases, I incremented the cost, then timed re-running the script. Of course, Python is an interpreted language, so the absolute times would be much lower if this were implemented in C, or assembly. Further, this was done on my aging T61 laptop. Running it on my 8-core i7 workstation with triple-channel DDR3 would show improved times. It not the times that are critical. What is critical is seeing the exponential back-off as the cost is increased.

Here is a table detailing my findings. Notice that the bcrypt cost increments by a single digit. Because it's binary exponential back-off, the times increase by a power of 2 at each iteration. I also adjusted the sha512crypt cost a little to more closely match the bcrypt timings, even though it's not a strict doubling of each cost value.

bcrypt sha512crypt
cost time iterations time
6 0.032 5,000 0.036
7 0.045 10,000 0.047
8 0.064 20,000 0.064
9 0.114 40,000 0.105
10 0.209 80,000 0.191
11 0.384 160,000 0.368
12 0.745 320,000 0.676
13 1.451 640,000 1.346
14 2.899 1,280,000 2.696
15 5.807 2,560,000 5.347
16 11.497 5,500,000 11.322
17 22.948 11,000,000 22.546
18 45.839 22,000,000 45.252
19 1:31.95 44,000,000 1:30.14
20 3:07.27 88,000,000 3:07.52

In the Python bcrypt implementation, the default cost is "10". For most modern GNU/Linux operating systems, when storing the user password in /etc/shadow with sha512crypt (yes, I didn't come up with the name), the default cost is 5,000 iterations. In both these cases, the cost can be adjusted. In the case of the Python bcrypt module, it's just passing the function with a numerical argument. In the case of GNU/Linux, it's editing PAM by adding "rounds=" to a config file.

As such, sha512crypt can be just as slow as bcrypt. Remember, we are trying to adjust for increased computing power that password crackers will have access to. In both cases, bcrypt and sha512crypt address that requirement.

So, if sha512crypt can operate with a cost parameter similar to bcrypt, and can provide that exponential back-off that we are looking for to slow down password brute force searching, then what's the point of bcrypt? Are there any advantages to running it? It turns out, there is, and I suspect this is a consequence of the design, and not something that was intentionally added.

What we would like is to prevent password crackers from using non-PC hardware on attacking the password database. SHA-2 functions, such as SHA-512, work very well on GPUs. SHA-2 functions work well on specialized hardware such as ASICs and FPGAs. As such, while we could make things slow for CPU or GPU crackers, those password crackers with specialized hardware would still have an upper hand on attacking the password database. Further, by addressing GPU cracking, and making it intentionally slow there, we make like more difficult for CPUs, which means hurting the honest user when trying to login to your web application. In other words, if I adjusted sha512crypt for GPU crackers, such that only 1,000 passwords per second could be achievable on a GPU, that might be a full second, or more, for the user logging into your CPU server. This may or may not be desirable.

So, how does bcrypt attack this problem? Part of the algorithm requires a lookup table stored in RAM that is constantly modified during execution. This turns out to work out very well on a standard PC where the CPU has exclusive access to RAM. This turns out to work out fairly poorly on a GPU, where the cores share the on-board memory, and each core must compete on the bus for access to those registers. As a result, any additional cracking speed is greatly minimized on a GPU when compared to the CPU. In other words, GPU password cracking with bcrypt isn't entirely effective either.

For SHA-2 functions, like SHA-512 however, this is not the case. SHA-2 functions use only 32-bit logic and arithmetic operations, which GPUs excel at. By using a GPU over a CPU for our sha512crypt function, a password cracker can get a couple to many orders of magnitude of additional cracking power.

So, the reason to use bcrypt isn't because "it's slow". The reason to use bcrypt is because "it's ineffective on GPUs".

A better alternative- scrypt
Unfortunately for bcrypt, however, due to its low memory requirement, bcrypt can be implemented in a field programmable gate array (FPGA) or custom ASICs. Understand that bcrypt was designed in 1999, when such specialized hardware had low gate counts, and was few and far bewteen. 15 years later, times have drastically changed. Bitcoin ASICS with SHA-256 FPGAs on board are common place. Hardware AES is common in CPUs and embedded systems. The fact of the matter is, these FPGAs, with their onboard, and fast RAM are well suited to bring bcrypt password cracking well into "fast" territory.

An alternative would be a solution that not only requires address registers to be constantly modified during algorithm execution, but to also exponentially bloat the memory requirement for the increased cost. scrypt addresses this shortcoming in bcrypt. Scrypt is another password key derivation function that was initially designed in 2009 by Colin Percival as part of the Tarsnap online backup service.

Scrypt has all of the advantages that bcrypt provides- baked in salt, exponential cost parameter, and ineffectiveness on GPUs, while also adding an exponential RAM requiremnt per the cost. Because of this RAM requirement, it is no longer cost efficient to build FPGAs with the necessary RAM.

Security, standards, and implementations
Scrypt is only 5 years young. This gives bcrypt a good 10 year head start. In terms of security, this is preferred. We want cryptographic functions that have withstood the test of time with cryptographers constantly attacking and analyzing their functions, primitives, and implementations. The longer it remains "unbroken", the more "secure" we deem the functions to be.

Bcrypt continues to be attacked and analyzed, and is showing no serious sign of weakness, 15 years later. This is good for the security of bcrypt. Scrypt however has seen less scrutiny, mostly due to its young age. However, it has been 5 years, and like bcrypt, no serious signs of weakness have been shown. By comparison, the SHA-2 family of functions was created in 2001, and has been scrutinized much more than bcrypt and scrypt combined, and also is not showing any serious signs of weakness. So, from a security standpoint, the SHA-2, bcrypt, and scrypt functions all seem to be fairly secure.

When looking at governing body standards, NIST has no paper on bcrypt or scrypt. They do recommend using PBKDF2 (another key derivation function (which I haven't explained here, but love)) for password storage, and NIST has standardized on SHA-2 for data integrity. Personally, while I like the ideas of bcrypt and scrypt, I would recommend sticking with the NIST recommendations with high iteration counts, as shown above. Until we see more standards boy interest in bcrypt and scrypt, IMO, you are taking a risk using them for password storage (less so for bcrypt than scrypt at least).

Finally, because of the newness of scrypt, there are less tools for its use in programming languages than bcrypt, and even more so for SHA-2. Further, most programming languages don't include either bcrypt or scrypt in their "standard library" of modules or functions, while SHA-2 is more generally found. And for those implementations, some 3rd party libraries are more trust worthy than others. Because you're dealing with password storage, it's critical you get this right.

Conclusion
While I love the algorithms behind bcrypt and scrypt, I've always advocated for using high iterative counts on the SHA-2 or PBKDF2 functions. Even further is the advocating of teaching people how to understand the concepts behind their password entropy, and improving their own online security. That is the weakest link, and needs the most work, IMO.

So, if you ask me, I'll tell you to use either PBKDF2 or SHA-2 with high iterative counts. Or, if you must absolutely use bcrypt, then I'll recommend that you use scrypt instead.

## Super Size The Strength Of Your OpenSSH Private Keys

In a previous post, about 18 months ago, I blogged about how you can increase the strength of your OpenSSH private keys by using openssl(1) to convert them to PKCS#8 format. However, as of OpenSSH verison 6.5, there is a new private key format for private keys, as well as a new key type. The new key type is ed25519. Without going into the details of the strengths of ed25519 over RSA, I do want to identify a new encryption method for your private keys.

In previous versions of OpenSSH, if you provided a passphrase to encrypt your private key, it was converted into a cipher key by first hashing it with MD5, then encrypting your private key. Unfortunately in this case, as usually the problem with all hashed password storage, MD5 is fast, fast, fast. While MD5 is a cryptographic one-way hashing function, it can fall victim to rainbow table attacks, as well as just plain old brute forcing. With a few GPUs, and the right software algorithm, it's not unheard of to try billions of passwords per second to try and achieve the correct hash, or in this case a cipher key.

Key derivation functions (KDFs) however, can be resource intensive, and slow. One in particular is bcrypt, which is very similar to the bcrypt one-way hashing function. With OpenSSH 6.5, when generating ed25519 keys, the bcrypt pbkdf is the default function for creating that cipher key based on your passphrase. To further protect you from brute force searching on your passphrase, ssh-keygen(1) will apply 16 rounds to the bcrypt pbkdf before creating the cipher key which is used to encrypt the private key on disk. On my ThinkPad T61, this takes approximately 1/3 of a second to complete all 16 rounds, or about 50 per second. This is a far cry from the millions I know my T61 can do with MD5.

However, this isn't even the bread and butter of the post: You can convert your existing keys to the new format with OpenSSH 6.5. This means your old DSA and RSA keys, and even the newer ECDSA keys, can all be converted to take advantage of the new format.

Further, you don't have to take the default 16 rounds of encrypting your key. Instead, you can increase that if you want to be a bit more paranoid. Suppose I wish to apply 100 rounds instead of the default 16- a factor of over 6x. To do this, for each of your private keys, run the following:

$ssh-keygen -o -p -a 64 -f id_rsa Enter old passphrase: Key has comment 'rsa w/o comment' Enter new passphrase (empty for no passphrase): Enter same passphrase again: Your identification has been saved with the new passphrase. At this point, it will take approximately 2 seconds on my T61 to complete the rounds, and encrypt the key. This can be verified whet creating an SSH agent, and adding my key to the agent: $ eval $(ssh-agent) Agent pid 17202$ ssh-add
Enter passphrase for /home/aaron/.ssh/id_rsa:
Identity added: /home/aaron/.ssh/id_rsa (/home/aaron/.ssh/id_rsa)

When adding my passphrase, it takes a full 2 seconds before returning to a shell on the remote server. Of course, feel free to increase the rounds count. 1000 rounds would take me a full 20 seconds. Probably not sufficient for day-to-day use while at work, but could be applicable in other cases.

When you look at your private keys, with the old version, the header would look something like this:

-----BEGIN DSA PRIVATE KEY-----
Proc-Type: 4,ENCRYPTED
DEK-Info: AES-128-CBC,DF7C541751D59241F15DA424506137CE

If you converted your key to PKCS#8 with openssl(1), then your headers would look something like this:

-----BEGIN ENCRYPTED PRIVATE KEY-----
(base64 output)

However, with the new OpenSSH key format, encrypted keys now look like:

-----BEGIN OPENSSH PRIVATE KEY-----
(base64 output)

With bcrypt as the new encrypted storage format, and the ability to adjust the number of rounds, as well as convert older keys, this is a big win for security. Well done OpenSSH team!

UPDATE: It should be noted that when using this new on-disk encrypted format, your OpenSSH private key will no longer be compatible with openssl(1), as previously, the private key was stored in PEM format. Further, using the "ed25519" key type means using the new format automatically, as openssl(1) does not support the ed25519 algorithm.

## Use /dev/random Instead Of /dev/null

While writing a shell script the other day, I was redirecting some output to /dev/null, as normal, when something dawned on me. Why don't I redirect my output to /dev/random instead? After all, both Linux random devices are writable by everyone on the system:

$ls -l /dev/*random crw-rw-rw- 1 root root 1, 8 Nov 13 15:14 /dev/random crw-rw-rw- 1 root root 1, 9 Nov 13 15:14 /dev/urandom Knowing what I know about the Linux cryptographic pseudorandom number generator (CSPRNG), I know that any bits put into the CSPRNG input pool are hashed with the SHA1 cryptographic hash function 512-bits at a time. This includes any data you redirect to it from the shell, as well as output from itself. When data is fed into the CSPRNG input pool, the RNG is reseeded. To understand this concept of seeding an RNG, let's assume for a moment that the only source of input for the RNG is its own output. If this were the case, we would only need a starting value to "seed" the RNG, then let it run by hashing its own digests. In this scenario, each digest is chosen deterministically, and if we know the input that seeded the RNG, we can predict all of its future outputs. Think of this scenario like a progress bar. For the SHA1 cryptographic hash, there are 2^160 possible unique digests. Theoretically, our RNG should be able to work through all 2^160 digests only once before starting over, provided there is enough time to do so (we know now this isn't the case, which means SHA1 has been weakened in terms of collision attacks). However, when you change the input by providing something other than the next digest in the queue, you change the next starting point of the RNG. It's as though you've "skipped" to a non-sequential location in your progress bar. Now, consider constantly reseeding your RNG. This is what your Linux system is actually doing. It's constantly processing timing events from disk IO, network packets, keyboard presses, mouse movements, etc. All these inputs get collected into an "entropy pool", which is then hashed with SHA1 512-bits at a time, as we mentioned earlier. This input changes the sequential ordering of the digest outputs, making the result unpredictable, and non-deterministic. So, when working on the shell, by redirecting your output to /dev/random, you reseed the CSPRNG, meaning you have changed the digest output ordering to something different than what it would have been had you not redirected those bits. In fact, the more you send data to the CSRNG, the more you reseed it, forever altering the path it takes on its "progress bar". Now, you may ask why not have some userspace PID running in the background that is always reseeding the CSPRNG? Sure, you can. In this case, I would recommend running Haveged on your system. Haveged will probe much more hardware events on the system than the default install will, and keep the entropy pool topped off at full. The CSPRNG will be constantly reseeded. However, for shell scripts, redirecting to /dev/random instead of /dev/null works. My only concern with redirecting to /dev/random would be bandwidth concerns. Doing a simple and crude benchmark comparing /dev/null to /dev/random, I get the following on my workstation: $ for I in {1..5}; do dd if=CentOS-6.5-x86_64-bin-DVD1.iso of=/dev/random; done
8726528+0 records in
8726528+0 records out
4467982336 bytes (4.5 GB) copied, 81.3842 s, 54.9 MB/s
8726528+0 records in
8726528+0 records out
4467982336 bytes (4.5 GB) copied, 76.4597 s, 58.4 MB/s
8726528+0 records in
8726528+0 records out
4467982336 bytes (4.5 GB) copied, 74.6036 s, 59.9 MB/s
8726528+0 records in
8726528+0 records out
4467982336 bytes (4.5 GB) copied, 75.4946 s, 59.2 MB/s
8726528+0 records in
8726528+0 records out
4467982336 bytes (4.5 GB) copied, 74.375 s, 60.1 MB/s
$for I in {1..5}; do dd if=CentOS-6.5-x86_64-bin-DVD1.iso of=/dev/null; done 8726528+0 records in 8726528+0 records out 4467982336 bytes (4.5 GB) copied, 59.325 s, 75.3 MB/s 8726528+0 records in 8726528+0 records out 4467982336 bytes (4.5 GB) copied, 56.5847 s, 79.0 MB/s 8726528+0 records in 8726528+0 records out 4467982336 bytes (4.5 GB) copied, 54.4541 s, 82.1 MB/s 8726528+0 records in 8726528+0 records out 4467982336 bytes (4.5 GB) copied, 56.0187 s, 79.8 MB/s 8726528+0 records in 8726528+0 records out 4467982336 bytes (4.5 GB) copied, 57.0039 s, 78.4 MB/s Seems I get slightly better throughput with /dev/null, which isn't surprising. So, unless you know you need the throughput, I would recommend sending your data to /dev/random over /dev/null. ## The Bitmessage Proof Of Work I've been on the Bitmessage network roughly since it was released. Maybe only a month or two later. One thing that has had me intrigued, although I've never really paid attnetion to ut until now, is Bitmessage's proof-of-work puzzle. A proof-of-work puzzle is a puzzle your computer solves to generally gain access to some resource. Usually, the intention is to ether prevent a denial of service attack, or to prevent spamming. In the case of Hashcash, that I have blogged about many times here, uses CPU stress to find a solution to a SHA1 digest. Its intent is to fight spam. The guided tour protocol is a proof-of-work puzzle to prevent denial of service attacks on a network or to a server. It forces every client, regardless of resources, to add network latencies by making roundtrip network connections to different servers, before reporting back with the solution to the requested resource. There are other proof-of-work puzzles that can require memory consumption before being granted access. The point is, a proof-of-work puzzle shows the requested resource, server, or destination that the client has spent valuable time solving a puzzle, proving they've done the necessary work. Mining Bitcoin works on this principle. In the case of Bitmessage, as of protocol 3, the proof-of-work puzzle is a CPU stress test that is based on the size of the message being sent. The idea is to prevent spam from hitting inboxes. Unfortunately, the main client PyBitmessage is written in Python, so its ability to solve the proof-of-work is far slower than a compiled C/C++ implementation. The proof-of-work is defined as follows: Where: nonceTrialsPerByte = 1000 (default difficulty set by the Bitmessage owner during key generation, and defined as '1') payloadLengthExtraBytes = 1000 (to add some extra weight to small messages) payload = embeddedTime + encodedObjectVersion + encodedStreamNumber + encrypted payloadLength = the length of payload, in bytes, + 8 (to account for the nonce which we will append later) TTL = the number of seconds in between now and the object expiresTime (default is 28 days, plus or minus five minutes) So, by default for most Bitmessage addresses, this equation can be roughly simplified as: which can be further simplified to: With a small payload, say 100 bytes in size, our target becomes very large: The largest the payload can be is 256 kilobytes. Thus, our target becomes much smaller: So, the larger the message, the smaller the target. Further, if you increase your difficulty on a new key from 1 to 2, then the "nonceTrialsPerByte" becomes 2000, instead of 1000. This drops the target to an even smaller number. This "target" value becomes the benchmark by which the difficulty of the proof of work is defined. Now that we have our target value, we must set a trial value, and try to find a number deterministically that becomes smaller than our target. We do this with the SHA512() cryptographic hashing function. First, we set "trialValue = 99,999,999,999,999,999,999", or about 226 million times larger than our target value could ever be. Then, we take the SHA512(payload), and set a counter to 0 (called a "nonce" in the protocol). Now we enter the following loop: while trialValue > target: nonce = nonce + 1 resultHash = SHA512(SHA512(nonce||initialHash)), where "||" is concatenation trialValue = the first 8 bytes of resultHash, converted to an integer As you can see, the larger our target is, the easier it will be to find a trial value that is smaller than the target. However, the smaller the target value becomes, the more difficult it will become to find a smaller trial value. That target decreases in value as the difficulty or the message length is increased. Suppose my target was "385,531,657,911", and the first 8 bytes of my SHA512 digest value in hexadecimal was "0a4b6ff992e295fa". The decimal value of this digest is "741,809,681,334,441,466". In our example, this number is larger than my target, so I'll need to increment my counter by 1, and try again. In fact, the largest my 8-byte digest value in hexadecimal can be, is "00000059c37a3eb6". In otherwords, this is a dynamic Hashcash system with a double SHA512 instead of a single SHA1. Initially, I thought this proof-of-work system was a bit strange. I couldn't understand why the core developer(s) didn't choose a simple implementation of Hashcash. If the idea is to prevent spam from hitting the network, then Hashcash in and of itself will suffice. However, by placing the core difficulty on the length of the message, you discourage large binary abuse, such as trading images, music, or movies across the network. That's what Bittorrent is for. Since the max message size is now 256 KB as of protocol version 3, even more so. But, spammers could still easily send links to nefarious sites, which would be nothing more than 100 bytes or so, and calculate the proof-of-work easily. So, while large messages might be discouraged, spam could still be a problem. This is where increasing the difficulty upon key creation is useful. So, I like that the proof-of-work system not only has small message spam fighting built into the system, I also like the fact that it discourages large message abuse as well. It seems, at least from what I've studied at this point, that the proof-of-work system for Bitmessage is well thought through, and fairly mature. However, until the default client is written in C/C++, I fear that the proof-of-work is targeted primarily at Python implementations, which are 30-50x slower than their C/C++ counterparts. So, we may still see message abuse on the network until this is addressed. ## Using The Bitmessage Storage Service While hanging out on the "privacy" channel on Bitmessage, someone sent the following: "You have no files saved. For instructions please send a message to BM-2cUqBbiJhTCQsTeocfocNP5WCRcH28saPU with the subject 'help'." This is actually pretty cool. No doubt should you call into question a faceless storage provider, but I thought I would play around with it. When you send 'help' to the address, you get the following information: "You send commands to the service via messages, the message subject is used to tell the server what you want to do. You are authenticated via the address you send from, so be sure to keep it safe. Below are the possible commands: LIST - Lists all the files you currently have saved to your account. NEW [file_title] - Save a new file to your account. e.g "NEW my top secret file", put the contents of your file into the message body. UPDATE [file_id] [new_file_name (optional) ] - Update an existing file saved to your account. The new_file_name is optional and is only used to rename your file. e.g "UPDATE a567f My even more top secret file" REMOVE [file_id] - Removes a file saved to your account. e.g "REMOVE a567f". TOP TIP!!! When sending the LIST command, type some random blurb into the message. If you send multiple LIST commands to the server in a short period of time the Bitmessage client see's your requests as duplicate messages and ignores them. The random blurb ensures the server always hears you." Intrigued, I started thinking about this. Due to the limitation of BM sending up to 256KB messages, your file can be no bigger than 256K, unless you setup a striped RAID array. Then you can chop up the file into multiple messages. However, the messages will likely be decrypted at the other end, and will very likely be stored unencrypted. So, they need to be encrypted, then converted into base64 before sending to the address. Initially, I figured GnuPG would work for this. But then I thought that it's not a good fit, because I lose plausible deniability. So, instead, I'll use the 'dm-crypt' module with the cryptsetup(8). First, I need my block devices. I'll setup some files with dd(1), then add them to loopback devices. Notice that I'm building the files from /dev/urandom. This is critical, so no metadata, including encryption boundaries, is leaked: $ dd if=/dev/urandom of=~/file1 bs=256k count=1
$dd if=/dev/urandom of=~/file2 bs=256k count=1$ ls -l file?
-rw-rw-r-- 1 user user 262144 Nov 29 07:57 file1
-rw-rw-r-- 1 user user 262144 Nov 29 07:57 file2
$sudo losetup /dev/loop1 ~/file1$ sudo losetup /dev/loop2 ~/file2
$losetup -l NAME SIZELIMIT OFFSET AUTOCLEAR RO BACK-FILE /dev/loop1 0 0 0 0 /home/user/file1 /dev/loop2 0 0 0 0 /home/user/file2  Because I want plausible deniability, I'll use cryptsetup(8) first, _then_ create the RAID array. If I create the RAID array first, the two files will reveal information that they belong to an array. I don't want any metadata leaked at all. $ sudo cryptsetup create aes-crypt-1 /dev/loop1
$sudo cryptsetup create aes-crypt-2 /dev/loop2$ ls -l /dev/mapper/aes-crypt-?
lrwxrwxrwx 1 root root 7 Nov 29 07:47 /dev/mapper/aes-crypt-1 -> ../dm-0
lrwxrwxrwx 1 root root 7 Nov 29 07:47 /dev/mapper/aes-crypt-2 -> ../dm-1


I can now create the RAID array, format it wxth ext2, and mount it. A couple notes: I'll first want to set the RAID chunk size to something low, otherwise I won't be able to put down a filesystem on the block device. So, I chose the minimum according to mdadm(8), which is 4KB. Then, when formatting with ext2, I'll only get a total of 64 inodes, which means a total of 64 files. I'm going to increase this to 256, so I can intentionally fragment the filesystem before putting down the data. I'll explain the reason for the intentional fragmentation in a second.

$sudo mdadm --create /dev/md0 --level 0 --raid-devices 4 --chunk 4 /dev/mapper/aes-crypt-{1,2}$ sudo mdadm --detail /dev/md0
/dev/md0:
Version : 1.2
Creation Time : Sat Nov 29 07:47:19 2014
Raid Level : raid0
Array Size : 464
Raid Devices : 2
Total Devices : 2
Persistence : Superblock is persistent

Update Time : Sat Nov 29 07:47:19 2014
State : clean
Active Devices : 2
Working Devices : 2
Failed Devices : 0
Spare Devices : 0

Chunk Size : 4K

Name : example:0  (local to host example)
UUID : 2bca1bf9:3af4a5d1:1989bb34:9b46bb9c
Events : 0

Number   Major   Minor   RaidDevice State
0     253        0        0      active sync   /dev/dm-0
1     253        1        1      active sync   /dev/dm-1


Now the formatting. Notice I'm changing the number of inodes. Also, we don't need to set aside any space for the root user: It's occupying precious disk space.

$sudo mkfs.ext2 -N 256 -m 0 /dev/md0$ sudo mount /dev/md0 /mnt
$ls /mnt lost+found$ df -h /mnt
Filesystem      Size  Used Avail Use% Mounted on
/dev/md0        427K   15K  412K   4% /mnt
$df -i /mnt Filesystem Inodes IUsed IFree IUse% Mounted on /dev/md0 256 11 245 5% /mnt  Now before putting data into the filesystem, I'm going to fill it with small files of random data, then remove every n-th file to create enough space to put down my data. This fill force my data to be intentionally fragmented. The reason for this is to avoid a snapshot attack on the files I'll be storing remotely. If I did not fragment the data, then as I update the filesystem, only small incremental changes will take place. This will allow the attack to "subtract" a previous filesystem iteration from the current, allowing them to know where my stored data resides, as well as possibly figuring out what it contains. Because we're talking about a 512 KB filesystem here, even encrypted, disk I/O isn't a concern. First, how big can each file be at a maximum? It appears to be 1721 bytes. $ echo '(412*1024)/245' | bc -l # 245 available inodes, 412K available space
1721.99183673469387755102


Next, create the files, each with 1721 bytes:

$for i in {1..245}; do sudo dd if=/dev/urandom of=/mnt/file$i bs=1721 count=1 2> /dev/null; done


Unfortunately, the filesystem filled up before completion. As such, we have empty files. So, we'll find those and remove them:

$find /mnt -empty -type f | wc -l 25$ sudo find /mnt -empty -type f -delete
$ls /mnt/file* | wc -l 204  Now, I'm ready to fragment the filesystem. I know that one file I want to copy is 8820 bytes in size. So I need to free up 6 non-contiguous files. according to my math, I need to free up every 34th file: $ echo '8820/1721' | bc -l
5.12492736780941313190
$echo '204/6' | bc -l 34.00000000000000000000$ sudo rm /mnt/file{$((34*1)),$((34*2)),$((34*3)),$((34*4)),$((34*5)),$((34*6))}
$df -h /mnt Filesystem Size Used Avail Use% Mounted on /dev/md0 427K 415K 12K 98% /mnt  I'm now ready to copy in my 8820-byte file: $ sudo cp ~/secret-file.txt /mnt
$df -h /mnt Filesystem Size Used Avail Use% Mounted on /dev/md0 427K 424K 3.0K 100% /mnt  Now I can tear everything down: $ sudo umount /mnt
$sudo mdadm --stop /dev/md0$ sudo cryptsetup close aes-crypt-1
$sudo cryptsetup close aes-crypt-2$ sudo losetup -d /dev/loop1
$sudo losetup -d /dev/loop2  Now I need to convert my two files to base64, copy, paste, and send the resulting output to BM-2cUqBbiJhTCQsTeocfocNP5WCRcH28saPU. The subject of the first message would be "NEW file1", while the subject of the second message would be "NEW file2". The body of each message would be the base64 output for file1 and file2. $ base64 file1 | tr -d '\n'; echo
yAFtXPy7NgHl5Q0ueJiZgjOlYyrocWaxcvA6CjKF0rNd10sTfNvMCmQrL0cA79oO0 ...(snip) ...

$base64 file2 | tr -d '\n'; echo T3GTNvwiewXnnOHTjITNpukyLz4d8iv8wl/JP0YjY0v5s5euF1qv4WwMv9Ejl9AsNMm5NXoK/hFQK ...(snip)...  Of course, sending 2 messages with 256 KB in size will take some time to calculate the PoW, so be prepared for that (I don't know why the PoW is what it is. It should have been a double SHA256() like Bitcoin. Meh) When you want to update your files, you wll need to build everything backup, except you won't "create" the RAID array, you'll "assemble" it, and you won't format your filesystem, obviously: $ sudo losetup /dev/loop1 file1
$sudo losetup /dev/loop2 file2$ sudo cryptsetup create aes-crypt-1 /dev/loop1
$sudo cryptsetup create aes-crypt-2 /dev/loop2$ sudo mdadm /dev/md0 --assemble /dev/mapper/aes-crypt-1 /dev/mapper/aes-crypt-2
$sudo mount /dev/md0 /mnt  You'll also need to move your file off the filesystem and rewrite all the random files, before copying the new data on. This is important, because we're trying to prevent a snapshot attack against our encrypted filesystem. So, on every commit to the BM storage service, as best as we can control, every bit on the filesystem needs to be changing. As the help mentions, sending frequent "LIST" commands can cause your message to be ignored, unless the body of the message includes unique random data for each message sent. So, the "TOP TIP" of sending random data in the body of the message is a good idea. As such, I'll run the following for each "LIST", practically guaranteeing that I'll see a unique string, and the query will succeed: $ dd if=/dev/urandom bs=2k count=1 2> /dev/null | rhash --sha3-512 -
4b5b19c9a8a61cf724cd710c6fd0c54edb46de2bfa55f2ec9a179a590b808993593d333f95dd6c607e9d366c385037cc0a600d262898e3b4c5be26479d3c962c  (stdin)


I'll copy and paste the result into the message body before sending.

I don't know that I'm ready to store sensitive data with it, but I do think it's pretty cool. The 256KB limit for the file size is pretty damning, so this isn't something to use for intense media storage, but instead, mostly for documents, such as password storage, browsing bookmarks, student grades, etc. It works, although it's definitely a niche product for a niche group. I wouldn't expect my wife to use this.

## Where Cryptographic Hashing Algorithms Fail

What Is A Cryptographic Hashing Algorithm?
Cryptographic hashing algorithms are one-way functions that produce a message digest that represents a given input. Because the keyspace is so astromically large, it should be practically infeasible to find a different input that represents the same digest. The input is typically referred to as the message while the output is typically referred to as the digest.

Cryptographic hashes usually hold to four main principles:

• Computing a digest should be easy for any given message (some would also say it should be fast).
• When changing the message, the digest should exhibit the "avalanche effect".
• It should be practically infeasible to find two messages that produce the same digest.
• It should be practically infeasible to produce the message for a given digest.

Of all of the many cryptographic hashing algorithms in the world today, the following list are still considered to be secure:

• RIPEMD-160, RIPEMD-256, RIPEMD-320
• SHA-224, SHA-256, SHA-384, SHA-512, SHA-512/224, SHA-512/256
• SHA3-224, SHA3-256, SHA3-384, SHA3-512, SHAKE128, SHAKE256
• Skein-256, Skein-512
• Whirlpool

SHA3 is a new standard adopted by the United States National Institude of Standards and Technology (NIST) which uses Keccak as the underlying function. Keccak uses the sponge contruction for creating its digests, which has some interesting properties that we'll look at in a second.

The Workhorse Of Cryptography
Many call cryptographic hashing algorithms the workhorse of cryptography, and rightfully so. These hashing functions are used for a great deal of things:

• Pseudo-anonymous user- Hash the username, and pseudonymize.
• Data deduplication- Hash the data blocks, and compare.
• Message authentication- Hash the message, and send.

Cryptographic hashes are used in a lot of ways. PGP fingerprints are nothing more than SHA-1 hashes of a timestamp. Similarly, OpenSSH fingerprints are MD5 results. ZFS uses SHA-256 for data and metadata integrity in its filesystem. Bitcoin uses double SHA-256 functions for the block chain, transactions, and mining. Cryptographic hashing algorithms are used EVERYWHERE. Unfortunately, developers keep reinventing the wheel, and don't pay attention to the problems that come with cryptographic hashing algorithms.

Failure #1- Raw Speed
While certainly an advantage, the speed at which cryptographic hashing algorithms creates the digest can also be a big drawback. Consider storing passwords to disk. Hopefully by now, developers know not to store the passwords in plaintext in the database. If not, get a new job. So, if we hash the password with say SHA-512, then according to our principles discussed at the beginning of this post, it should be infeasible to find the password (message) that produced the hash stored on disk.

If this hashed password database is leaked, the password cracker will be attempting to hash as many messages as possible to retrieve the given digest. Typically, this is done using custom word lists, hashing each word, and comparing it to what is stored in the password database. The faster the password cracker can go, the sooner they can recover the digests. On my wimpy laptop, using the CPU alone, I can hash approximately 1 million passwords per second with SHA-512. Using cheap GPU clusters, it's feasible to go upwards of 50 billion passwords per second, for basement-type miscreants living in their mother's basement.

Possible Speed Fix #1- Key Streching
One way around brute forcing passwords is a concept called key stretching. This is the concept where I would take the hash of the hash of the message. In other words, the message is recursively hashed. So, if my password is hashed 5,000 times, then it would take the password cracker 5,000 times more work to get to the original message. So, my wimpy laptop is reduced to going through only 200 passwords per second, instead of 1 million. Anything to slow the attackers down, is a good thing.

A password key stretched twice would look like this, where "H" is the hashing algorithm:

H(H(password))

For example, taking the SHA-256 of "password" twice gives the following hexadecimal output:

SHA-256(SHA-256("password")) = 73641c99f7719f57d8f4beb11a303afcd190243a51ced8782ca6d3dbe014d146

Possible Speed Fix #2- Key-based Derivation Functions
Another alternative, and there is a good following for this on the Internet, is using key-based derivation functions. Without getting into too much detail here (I can save this for another post), key-based derivation functions like PBKDF2, bcrypt(), and scrypt() are slow, slow, slow, and very resource intensive. As such, this greatly limits the speed in which the password cracker can recover my password, much like key stretching previously mentioned.

Failure #2- Hashed Messages
Unfortunately, there's this thing called rainbow tables. Rainbow tables simply put, are large databases with message to digest mappings. Rainbow tables are built by hashing messages of all types, and storing both the message and the digest in the database. Then, all it takes is to look up the digest in the database. If it exists, you have the message that produced it. In this case, you have the password.

Thankfully, rainbow tables are usually bloated in size, typically several terabytes. The standard password cracker isn't going to dedicate 10 TB of disk just to store rainbow tables for MD5, SHA-1, and NTLM hashes. However, due to the one-to-one relationship (in practice) of messages to digests, rainbow tables are a very effective way to recover messages from digests only.

Possible Message Fix- Salting Hashed Messages
The number one way to thwart rainbow table attacks, is to prepend a cryptographic nonce, also called a "salt" to the password. The larger the space of the salt, the more possibilities the password could be when stored to disk. For example, if the salt is chosen from a base64 set of characters, and the salt is 8 characters long, then the salt's keyspace is 64^8 or 281,474,976,710,656 possible salts. This means that the rainbow table must store 281,474,976,710,656 digests for every message. As mentioned, rainbow tables are already several TB in size. Now they must be 281 trillion times larger.

A salted message looks like this, where "H" is the hashing algorithm, and "||" indicates concatenation:

H(salt || message)

For example, if my salt was "FhuVTwA710", and the password is "password", then taking the SHA-256 would give the following output:

SHA-256("FhuVTwA710" || "password") = 4c700717f3e5eb92a872362108f2f716d2ff179ea94d1f10853a50e181a43663

Failure #3- Salted Message Authentication Codes
Message authentication codes, or MACs for short, are a way to send a message between two parties that have previously authenticated. The idea is to send a message from Alice to Bob, such that when Alice's message reaches Bob, he knows whether or not if the message has been tampered with, based on if the hash of the message produces the same hash as what Alice sent. Alice and Bob must have met previously, and agreed on a key that they would use as a "salt" for each message they send.

In other words, their MAC setup looks like this, again where "H" is the hashing algorithm, "key" is the key both Alice and Bob agreed on, and "message" is what is being sent:

H(key || message)

For example, suppose the message was "The quick brown dog jumps over the lazy dog.", and the agreed-upon key is "J6n^,1=/#;RCA5bC10". Then our SHA-256 MAC would look like this:

SHA-256("J6n^,1=/#;RCA5bC10" || "The quick brown dog jumps over the lazy dog.") = 10c98713fbaa574ad9b2ffc053662c89f6fd387a8a350f6324b966d36936d1d3
MAC = ("The quick brown dog jumps over the lazy dog.", "10c98713fbaa574ad9b2ffc053662c89f6fd387a8a350f6324b966d36936d1d3")

Alice would send the full MAC (both the message and its digest). The idea is if the message changes, the digest will change, and as such, Bob will be able to detect tampering. Eve, a malicious 3rd party, should not be able to send fraudulent message to Bob, because Eve does not know the key Alice and Bob agreed on.

However, SHA-256 uses the Merkle–Damgård construction. Merkle–Damgård is iterative in nature- IE: block-by-block. So, while Eve may not know the private key, she was able to grab both the message and its digest in transit. So, all Eve needs to do is concatenate additional data after the message, know the length of the key, calculate the new digest, and ship this message to Bob. This is known as the length-extension attack.

In other words, if you know what the digest of H(key || message) is, then you can easily determine what H(key || message || ANYTHING) is. So to compromise the MAC system, you just need to see one message and digest MAC, then you can impersonate the sender from there on out. This should bother you.

Possible MAC Fix #1- The Secret Suffix
At this point, you may be thinking to switch the order of the key and the message, or H(message || key). This is known as the "secret suffix", and it is certainly a very valid fix to address the length-extension attack. However, it comes with a large assumption that you are making about the underlying cryptographic hashing algorithm.

That assumption is that the algorithm does not have any known digest collisions. In other words, if two messages m1 and m2 can be found to produce the same digest, then the attacker will be able to use this collision on the sent message, to use a different start message, putting the hashing algorithm state in the same starting position, and thus producing the same MAC, even though the key is still unknown.

Currently, both SHA-1 and SHA-2 do not have any known collisions, but it's only a matter of time.

Possible MAC Fix #2- The Enveloped MAC
Another possible solution is to envelope the message around the key, or H(key || message || key). This requires that the attacker know the key length, in order to identify the starting point of the message, and thus be able to forge a valid MAC. While more secure than the secret suffix, there has been some research in this area that suggests weaknesses to this approach, even when two different keys are used.

Possible MAC Fix #3- SHA-224, SHA-384, or SHA-3 (Keccak)
An actual solid fix is to use SHA-3, which is based on Keccak, for your hashing algorithm. SHA-3 is not vulnerable to the length-extension attack, and as such, can be used for a MAC. This is due to the fact that SHA-3 is based on the sponge construction, which is not an iterative block-by-block compression function like Merkle–Damgård is. With the existence of SHA-3, we may not need to worry about the next section for sending MACs.

Also, SHA-224 and SHA-384 are not vulnerable to the length-extension attack, due to the internal truncation of the internal state, whereas MD5, SHA-1, SHA-256, and SHA-512, among others, output the entire state.

Solution- The Hash-based Message Authentication Code (HMAC)
HMAC securely addresses the MAC length-extension attack. It does so by taking the hash function twice in a deterministic manner, first with an inner key, and again with an outer key. The algorithm is shown here:

1. Prepare the key:
1. If the key is less than the hashing algorithm block size, then append zeros to the key until it is the same size as the hashing algorithm block size.
2. If the key is greater than the hashing algorithm block size, then take the hash of the key, using the digest as the new key.