Stung by revelations of ubiquitous surveillance and compromised software, the internet’s engineers and programmers ponder how to fight back
SECURITY guards (at least the good ones) are paid to be paranoid. Computer-security researchers are the same. Many had long suspected that governments use the internet not only to keep tabs on particular targets, but also to snoop on entire populations. But suspicions are not facts. So when newspapers began publishing documents leaked by Edward Snowden, once employed as a contractor by America’s National Security Agency (NSA), the world’s most munificently funded electronic spy agency, those researchers sat up.
They were especially incensed by leaks published in September by the Guardian and theNew York Times, which suggested that American spooks (with help from their British counterparts) had been working quietly for years to subvert and undermine the cryptographic software and standards which make secure communication over the internet possible. “At that point”, says Matthew Green, a cryptographer at Johns Hopkins University, “people started to get really upset.”
On November 6th a meeting in Vancouver of the Internet Engineering Task Force (IETF), an organisation which brings together the scientists, technicians and programmers who built the internet in the first place and whose behind-the-scenes efforts keep it running, debated what to do about all this. A strong streak of West Coast libertarianism still runs through the IETF, and the tone was mostly hostile to the idea of omnipresent surveillance. Some of its members were involved in creating the parts of the internet that spooks are now exploiting. “I think we should treat this as an attack,” said Stephen Farrell, a computer scientist from Trinity College, Dublin, in his presentation to the delegates. Discussion then moved on to what should be done to thwart it.
We have the technology
As a sort of council of elders for the internet, the IETF has plenty of soft power. But it has no formal authority. Because its standards must be acceptable to users and engineers all over the world, it works through a slow process of consensus-building. New standards, guidelines and advice take months or years to produce.
Others, equally offended by the intelligence agencies’ activities, prefer not to wait, and are simply getting on with the job of trying to restore confidence in online security. As Bruce Schneier, a leading cryptographer, told the conference, it seems spies cannot actually break most cryptographic codes. Instead they try to work around them. One way is to subvert the standards and software which implement cryptography. That is possible because, besides trying to defeat the cryptographic efforts of others, the NSA also helps produce ciphers for Americans to use. Those same cryptographic standards are then employed all over the internet.
Researchers have therefore been warning users against employing anything that might have been tampered with. RSA Security, a big maker of encryption software, has advised its customers to stop using a random-number generator widely believed to have been fiddled with by the spooks to make its output predictable (random numbers are a crucial component of any cryptographic scheme, but are notoriously hard to produce on a deterministic machine such as a computer). And a group of Brazilian mathematicians has published a new set of codes for use with elliptic-curve cryptography, a novel scrambling technique that has been championed by the NSA. Anyone worried by the provenance of NSA-supplied curves is free to use these new ones instead.
Even America’s government is getting in on the act. The credibility of its National Institute of Standards and Technology, which sets American cryptographic standards with the help of the NSA, has been dented by Mr Snowden’s revelations.