RSA BSAFE is a FIPS 140-2 validated cryptography library offered by RSA Security. From 2004 to 2013 the default random number generator in the library contained an alleged kleptographic backdoor from the American National Security Agency (NSA), as part of its secret Bullrun program.
The SSL-C library is an SSL toolkit in the BSAFE suite. It was originally written by Eric A. Young and Tim J. Hudson, as a fork of the open library SSLeay, that they developed prior to joining RSA. Like SSLeay, SSL-C supported SSLv2, SSLv3, TLSv1; while it also supports X.509v1 and X.509v3. SSL-C was first released in 1999. The algorithm was one of several available in the libraries that users were able to choose from.
Dual_EC_DRBG random number generator
From 2004 to 2013, the default cryptographically secure pseudorandom number generator (CSPRNG) in BSAFE was Dual_EC_DRBG, which contained an alleged backdoor from NSA, in addition to being a biased and slow CSPRNG. The cryptographic community had been aware that Dual_EC_DRBG was a very poor CSPRNG since shortly after the specification was posted in 2005, and by 2007 it had become apparent that the CSPRNG seemed to be designed to contain a hidden backdoor for NSA, usable only by NSA via a secret key. In 2007, Bruce Schneier described the backdoor as "too obvious to trick anyone to use it." The backdoor was confirmed in the Snowden leaks in 2013, and it was insinuated that NSA had paid RSA Security US$10 million to use Dual_EC_DRBG by default in 2004, though RSA Security denied that they knew about the backdoor in 2004. The Reuters article which revealed the secret $10 million contract to use Dual_EC_DRBG described the deal as "handled by business leaders rather than pure technologists". RSA Security has largely declined to explain their choice to continue using Dual_EC_DRBG even after the defects and potential backdoor were discovered in 2006 and 2007, and has denied knowingly inserting the backdoor.
So why would RSA pick Dual_EC as the default? You got me. Not only is Dual_EC hilariously slow – which has real performance implications – it was shown to be a just plain bad random number generator all the way back in 2006. By 2007, when Shumow and Ferguson raised the possibility of a backdoor in the specification, no sensible cryptographer would go near the thing. And the killer is that RSA employs a number of highly distinguished cryptographers! It's unlikely that they'd all miss the news about Dual_EC.— Matthew Green, cryptographer and research professor at Johns Hopkins University, A Few Thoughts on Cryptographic Engineering (From after the backdoor was confirmed, but before the $10 million secret deal was revealed by Reuters.)
As a cryptographically secure random number generator is often the basis of cryptography, much data encrypted with BSAFE was not secure against NSA. Specifically it has been shown that the backdoor makes SSL/TLS completely breakable by the party having the private key to the backdoor (i.e. NSA). Since the US government and US companies have also used the vulnerable BSAFE, NSA can potentially have made US data less safe, if NSA's secret key to the backdoor had been stolen. It is also possible to derive the secret key by solving a single instance of the algorithm's elliptic curve problem (breaking an instance of elliptic curve cryptography is considered unlikely with current computers and algorithms, but a breakthrough may occur).
Extended Random TLS extension
"Extended Random" was a proposed extension for the Transport Layer Security (TLS) protocol, submitted for standardization to IETF by an NSA employee, although it never became a standard. The extension would otherwise be harmless, but together with the Dual_EC_DRBG, it would make it easier to take advantage of the backdoor.
The extension was previously not known to be enabled in any implementations, but in December 2017, it was found enabled on some Canon printer models, which use the RSA BSAFE library, because the extension number conflicted a part of TLS version 1.3.
On November 25, 2015, RSA announced End of Life (EOL) dates for BSAFE. The End of Primary Support (EOPS) was reached on January 31, 2017, and the End of Extended Support (EOXS) was originally January 31, 2019, but has now been extended for some versions until January 31, 2022. During Extended Support, only the most severe problems will be patched.
- Menn, Joseph (December 20, 2013). "Exclusive: Secret contract tied NSA and security industry pioneer". San Francisco. Reuters. Retrieved December 20, 2013.
- Simson Garfinkel, Gene Spafford (2002). Web Security, Privacy & Commerce. O'Reilly. p. 114. ISBN 0596000456.
- Ivan Ristic (2013). OpenSSL Cookbook: A Guide to the Most Frequently Used OpenSSL Features and Commands. Qualys. p. 1. ISBN 9781907117053.
- RSA Data Security (1999). "RSA Introduces BSAFE SSL-C for Worldwide Markets". PR Newswire.
- RSA Australia (30 March 1999). "Check Point Software Licenses RSA BSAFE™ SSL-C for its Market-Leading VPN". Check Point Software. Archived from the original on 1 December 2010.
- RSA (2013). "RSA Response to Media Claims Regarding NSA Relationship". RSA.
- Matthew Green (September 20, 2013). "RSA warns developers not to use RSA products". A Few Thoughts on Cryptographic Engineering.
- Bruce Schneier. "The Strange Story of Dual_EC_DRBG".
- "We don't enable backdoors in our crypto products, RSA tells customers". Ars Technica.
- Eric, Rescorla; Margaret, Salter (2 March 2009). "Extended Random Values for TLS". IETF draft standard.
- Menn, Joseph (31 March 2014). "Exclusive: NSA infiltrated RSA security more deeply than thought - stu". Reuters.
- Green, Matthew (19 December 2017). "The strange story of "Extended Random"". Cryptographic Engineering blog.
- RSA (November 25, 2015). "RSA announces End of Life (EOL) dates for RSA BSAFE". RSA. Retrieved October 3, 2018.
- RSA (June 20, 2018). "RSA announces support extension for some of the BSAFE® product suite". RSA. Retrieved October 3, 2018.
- RSA BSAFE - RSA Security