recentpopularlog in

mcherm : cryptography   234

« earlier  
Latacora - The PGP Problem
Don't use PGP -- it was designed before the modern era in encryption and it's security sucks while it's usability is famously horrible.
cryptography  via:HackerNews  programming 
9 weeks ago by mcherm
The World’s Oldest Blockchain Has Been Hiding in the New York Times Since 1995 - Motherboard
A hash-certified chain has been running with timestamps in the New York Times since 1995.
cryptography  blockchain  nytimes  security 
august 2018 by mcherm
A riddle wrapped in a curve – A Few Thoughts on Cryptographic Engineering
This is a 2005 article positing a conspiracy theory that the NSA discovered a major weakness in elliptic curve cryptography.
cryptography  NSA  via:HackerNews 
august 2018 by mcherm
Myths about /dev/urandom [2uo]
A long and detailed piece about why you shouldn't block waiting for "more entropy" for your random number generator.
cryptography  security  programming  via:HackerNews 
august 2018 by mcherm
Exclusive: NSA encryption plan for ‘internet of things’ rejected by international body – WikiTribune
The US has pushed for certain encryption algorithms to become the new standard for internet of things. The international standards body rejected it because they suspect the NSA will attempt to insert back doors or breakable algorithms.
cryptography  security  standards  NSA  via:boingboing 
april 2018 by mcherm
Four cents to deanonymize: Companies reverse hashed email addresses
Lots of folks pass around lists of hashes of email addresses. These aren't secure because only a few billion email addresses exist and all can be hashed cheaply. Use salt (unless your goal is uniqueness testing, in which case don't share the list).
security  cryptography  email  via:boingboing 
april 2018 by mcherm
Cryptographic Right Answers — » Latacora
Current (2018) good advice on which specific cryptography algorithms to use.
cryptography  via:HackerNews 
april 2018 by mcherm
Monero Privacy Protections Aren’t as Strong as They Seem | WIRED
Monero is viewed as a cryptocurrency with untraceable transactions, but it merely mixes real and fake transactions on the blockchain and the real ones can often be identified.
via:HackerNews  wired  cryptocurrency  privacy  cryptography 
march 2018 by mcherm
Attacking Merkle Trees with a Second Preimage Attack – flawed.net.nz
Generate data that gives the same hash as a known value and you can undermine a merkle tree. Use some known structure and/or track depth to protect against this attack.
cryptography  security  via:HackerNews 
march 2018 by mcherm
Falling through the KRACKs – A Few Thoughts on Cryptographic Engineering
The WPA2 WiFi protocol was written years ago and was written years ago and was"proved" secure, yet a flaw was just found. Why? Blame IEEE for making the standards hard to access and poorly written; and the proof was done on two pieces in isolation instead of the whole system or better yet the actual code.
security  standards  IEEE  cryptography  via:HackerNews 
october 2017 by mcherm
Mastercard Internet Gateway Service: Hashing Design Flaw – Tinyhack.com
This security researcher identified a flaw in Mastercard's hashing and then identified a flaw in their fix as well. The cryptography of it is explained quite clearly. Lesson: use delimiters correctly.
security  payments  cryptography  via:HackerNews 
september 2017 by mcherm
How I implemented my own crypto
He wrote a crypto library and explains the level of analysis necessary to quash all C bugs as well as the difficulties in efficiently implementing published cryptography primitives.
cryptography  personal_net  programming  via:HackerNews 
august 2017 by mcherm
JWT (JSON Web Tokens) is a Bad Standard That Everyone Should Avoid - Paragon Initiative Enterprises Blog
Points out that a client can request the "none" encryption algorithm. Wait, what? Why can the client request anything? If libraries are accepting "none" and other broken encryption schemes, then THAT is the bug.
JavascriptWebTokens  security  cryptography  standards  via:HackerNews 
march 2017 by mcherm
MD5 considered harmful today
MD5 collisions were proven (in 2008) to actually be exploitable to create fake trusted SSL certs.
cryptography  security  ssl 
february 2017 by mcherm
SHA-1 collisions have now been demonstrated
We knew this was coming, but I thought it would take a lot longer.
cryptography  via:reddit  security 
february 2017 by mcherm
CIA Director John Brennan Pretends Foreign Cryptography Doesn't Exist - Schneier on Security
CIA director "perhaps a bit misleading" in testimony to Congress. Sadly, this is to be expected.
cryptography  politics  via:HackerNews 
june 2016 by mcherm
Padding oracles and the decline of CBC-mode cipher suites
Why CBC mode is unsafe, explained very carefully and clearly along with some history.
cryptography  security  ssl  via:HackerNews  history 
february 2016 by mcherm
HTTPS provides more than just privacy
A list of good reasons why your site you should use https instead of just HTTP.
internet  ssl  cryptography  via:HackerNews 
january 2016 by mcherm
Storing Passwords in a Highly Parallelized World · Homepage of Hynek Schlawack
The correct algorithm to use for encrypting passwords (Argon2i, beats bcrypt, which itself beats SHA-1 which beats md5 which beats rot26). Also a Python library for using it.
cryptography  security  python  via:HackerNews 
january 2016 by mcherm
Crypto is For Everyone—and American History Proves It | Electronic Frontier Foundation
An article by the EFF documenting how the US fou ding fathers made extensive use of cryptography.
via:reddit  eff  cryptography  history 
november 2015 by mcherm
How to Protect Yourself from NSA Attacks on 1024-bit DH | Electronic Frontier Foundation
EFF tells us how to avoid NSA listening now that we realize they've probably broken 1024-bit Diffie-Hellman key exchange for certain commonly used primes. Basically, they said use 2048 bit encryption instead.
eff  security  ssl  nsa  privacy  cryptography 
october 2015 by mcherm
1Password Leaks Your Data
1password's older (still standard) format puts the metadata (URL, name, notes, etc) in plaintext. It also encourages you to post the file online.
security  cryptography  via:reddit 
october 2015 by mcherm
How is NSA breaking so much crypto?
A reasonable sounding theory of how the NSA spies on even encrypted traffic off internet backbones.
security  cryptography  nsa  privacy  surveillance  via:reddit 
october 2015 by mcherm
Base 64 without padding
Burried in another RFC is how to use base64 without padding.
standards  cryptography  base64 
september 2015 by mcherm
CynoSure Prime: How we cracked millions of Ashley Madison bcrypt hashes efficiently
Ashley Madison used bcrypt so their passwords should have been unbreakable. But they ALSO used md5, so they were cracked.
security  cryptography  via:HackerNews 
september 2015 by mcherm
Ethan Heilman — A Brief History of NSA Backdoors.
A list of known big backdoors to cryptography put in by the NSA.
via:HackerNews  security  cryptography  nsa  history 
july 2015 by mcherm
A Few Thoughts on Cryptographic Engineering: A history of backdoors
The government has a history of trying "government access crypto" and it has been a series of miserable failures.
cryptography  security  privacy  via:HackerNews 
july 2015 by mcherm
A proposal to keep the amounts private in bitcoin
Roughly, provide proofs that the amounts add up without revealing the actual amounts.
cryptography  bitcoin  privacy  via:HackerNews 
june 2015 by mcherm
dpr » udp and me
How UDP got "invented" and some of the compromises made in TCP because they knew the network "would never get that big". Also no end-to-end crypto because the NSA said not to.
history  internet  networking  via:HackerNews  security  cryptography  nsa 
may 2015 by mcherm
NSA in P/poly: The Power of Precomputation
A paper has now been published that proves the following is possible (and likely being done by the NSA): pre-compute tables for a handful of prime numbers and thus break Diffy-Hellman in the real world (where the same few primes are almost always used).
cryptography  security  via:ScottAaronson 
may 2015 by mcherm
Moxie Marlinspike >> Blog >> The Cryptographic Doom Principle
Two very subtle protocol vulnerabilities, and a principle that avoids them. The principle is to verify the MAC as the very first thing that happens in a protocol.
cryptography  security  bugs  via:HackerNews  MoxieMarlinspike 
april 2015 by mcherm
Critical vulnerabilities in JSON Web Token libraries
Aaarg! I would have known better than to fall for something THIS dumb. Why am I asked to use this library?
security  cryptography  via:HackerNews 
april 2015 by mcherm
Moxie Marlinspike >> Blog >> GPG And Me
Mozie Marlinspike says GPG is just so difficult to use that we should pretend it doesn't even exist and start over.
security  privacy  cryptography  via:HackerNews 
february 2015 by mcherm
Life in a post-database world: using crypto to avoid DB writes
On the use of encrypted cookies passed to the client instead of storing temporary data in the database. Password resets are the canonical example.
security  programming  cryptography  via:HackerNews 
february 2015 by mcherm
Towards the Perfect Coin Flip: The NIST Randomness Beacon | Hackaday
NIST is generating public random numbers published on a fixed schedule.
via:HackerNews  random  cryptography 
december 2014 by mcherm
Using Freenet
hey, at least ONE person is out there trying to use freenet.
freenet  cryptography  via:HackerNews 
december 2014 by mcherm
Keeping Secrets — STANFORD magazine — Medium
About the publication of Diffie-Hellman: the NSA objected, but Stanford offered to support them legally.
history  cryptography  politics  via:HackerNews 
november 2014 by mcherm
81% of Tor users can be de-anonymised by analysing router information, research indicates
Evil entry port + traffic monitoring at exit and with colluding server means you can track most Tor traffic.
tor  security  cryptography  privacy  via:reddit 
november 2014 by mcherm
CREAM: the scary SSL attack you’ve probably never heard of
Just a talk about timing attacks on SSL to extract keys by closely observing the response time of code that is not constant-time.
cryptography  security  ssl 
november 2014 by mcherm
Why Google is Hurrying the Web to Kill SHA-1
Nice writeup of the situation intended for a non-technical audience.
cryptography  google  ssl  via:HackerNews 
september 2014 by mcherm
Google Online Security Blog: Gradually sunsetting SHA-1
Googly has a plan for Chrome to start gradually trusting certs signed by SHA-1 less and less to persuade people to move off them.
security  cryptography  ssl  google  chrome  via:HackerNews 
september 2014 by mcherm
Can Internet Security Ever Work?
An insightful essay on the difficulty of setting up authorities for identifying people (necessary for full security in a world with public key cryptography).
cryptography  security  via:HackerNews 
may 2014 by mcherm
ImperialViolet - No, don't enable revocation checking
SSL cert revocation lists are too big to store with every browser. Checking every time is so unreliable that browsers don't treat failures as fails.
security  ssl  cryptography  internet  via:HackerNews 
april 2014 by mcherm
Polypasshash
Clever idea. Don't store the hash of the passwords, use Shamir secret sharing and store a single point on the line. The server does NOT know the line, and at startup it can't validate passwords until after several people have given correct passwords. Probably can be beaten by opening a few accounts on the target system.
security  cryptography  via:slashdot 
april 2014 by mcherm
Kraken Passes Cryptographically Verifiable Proof of Reserves Audit
Now THIS is difficult to do with anything OTHER than Bitcoin. A company proves they have reserves sufficient to cover all of their accounts, and does so in a cryptographically reliable way that nevertheless does not reveal information about their accountholders.
cryptography  bitcoin  banking  via:reddit 
march 2014 by mcherm
Myths about /dev/urandom
Random number generation: why you (normally) do NOT have to wait to collect enough entropy.
math  cryptography  random  via:HackerNews 
march 2014 by mcherm
Bitcrypt broken - Cassidian CyberSecurity blog
Their friend was attacked by ransomware and they found a bug in it's cryptography (used too short a key) so they were able to overcome it.
cryptography  bug  via:HackerNews 
february 2014 by mcherm
Why is quality of pseudorandom number generators important? - Super User
A wonderful answer to the question "why good random number generators?". The answer shows how to cheat at poker if they aren't good.
EricLippert  random  programming  security  cryptography  math  via:HackerNews 
february 2014 by mcherm
How the Bitcoin protocol actually works | DDI
A really good explanation of how Bitcoin works, motivating each step of the design.
bitcoin  cryptography  via:BruceSchneier 
january 2014 by mcherm
Reverse engineering my bank's security token | Thiago Valverde
He reversed engineered his bank's security token generator. Turns out it was well-designed. Also, anything running on a client device can always be reverse engineered.
security  banking  cryptography  programming  via:reddit 
january 2014 by mcherm
Exclusive: Secret contract tied NSA and security industry pioneer
A claim (I'd need to see better evidence) that RSA accepted a bribe from the NSA to use a flawed random number generator as the default in their software.
nsa  rsa  evil  cryptography  security  via:reddit 
december 2013 by mcherm
Crypto Fails — Telegram's Cryptanalysis Contest
Analyzing a bad crypto system, he explains why modern cryptography must protect against attackers who can submit known plaintext and can freely modify encrypted messages and submit them for decryption.
cryptography  security  via:HackerNews 
december 2013 by mcherm
Acoustic cryptanalysis
A devastating, nearly unbelievable side channel attack. They can break RSA by listening to the SOUND of the CPU decrypting chosen data. And they can do this on real hardware (like a normal laptop) using just a CELL PHONE!
security  cryptography  sidechannel  via:reddit 
december 2013 by mcherm
Visual Cryptography
Use cryptographic secret sharing to hide data in some images that can only be found by combining them.
cryptography  images  secretsharing  via:reddit 
november 2013 by mcherm
In Lavabit Appeal, U.S. Doubles Down on Access to Web Crypto Keys | Threat Level | Wired.com
Effectively, the US government is arguing in the lavabit case that they have the right to demand encryption keys from anyone with a subpoena.
lavabit  privacy  security  law  cryptography  snowden  via:HackerNews  wired 
november 2013 by mcherm
Satoshi's Genius: Unexpected Ways in which Bitcoin Dodged Some Cryptographic Bullets – Bitcoin Magazine
Bitcoin did several things surprisingly well. It's numbers are always less than MAX_FLOAT (so it works in languages with just one number type), it publishes a hash of the public key not the key so it's resistant to quantum computers, and it picked an elliptic curve that's clearly not compromised by the NSA.
security  cryptography  crypto  bitcoin  via:HackerNews 
october 2013 by mcherm
1Password and The Crypto Wars | Agile Blog
Why you can trust us: our design prevents us from having access to the data and we have employees in several different countries who would tell if we undermined that design.
security  privacy  snowden  cryptography  via:HackerNews 
september 2013 by mcherm
The NSA’s work to make crypto worse and better
Details of how the NSA put a back door into a crypto standard. In 2006.
ArsTechnica  via:ArsTechnica  nsa  cryptography  security  privacy 
september 2013 by mcherm
N.S.A. Foils Much Internet Encryption - NYTimes.com
Based on Snowden's documents, the New York Times reports that that NSA has cracked nearly all encrypted communications on the internet, sometimes by obtaining keys, sometimes by forcing vendors to put in back doors, and sometimes by influencing standards bodies to create crackable algorithms.
nytimes  via:HackerNews  security  nsa  cryptography  privacy  snowden 
september 2013 by mcherm
Encryption is less secure than we thought - MIT News Office
When calculating entropy of messages for determining the risk of breaking a code, one should not use the average entropy but rather a form of the "worst case" because that's all the attacker needs. This is still exponentially hard, but exponentially less than we used to think.
math  cryptography  via:HackerNews 
august 2013 by mcherm
« earlier      
per page:    204080120160

Copy this bookmark:





to read