A new computer chip will help tackle one of the Web’s weak spots—the fact that most data is exchanged without any protection against hackers or eavesdroppers.
For some communications, such as credit card payments and online banking transactions, it is standard to encrypt the information that users and websites send each other. But most online activity is completely unprotected, largely because encrypting communications requires extra work from Web servers and software, which is costly to implement.
Search queries and social media updates, for example, are almost exclusively sent in forms easily read by a third party snooping on Web traffic. Listening in to Web traffic can be as simple as using the same Wi-Fi network as the target, as Ashton Kutcher found when his Twitter account was hijacked at the TED conference earlier this year, by means of a Firefox add-on called Firesheep.
A microchip developed by semiconductor design company Cavium could allow much more—perhaps even all—Web traffic to be encrypted, by reducing the cost of implementing encryption. Cavium’s Nitrox III chip is designed to be installed in data centers that serve up Web pages and manage Web apps. It’s specialized design is extremely fast and efficient at the mathematical calculations underpinning the encryption that secures Web sites that use the protocol SSL. Sites secured this way have Web addresses that start with HTTPS, instead of HTTP.
When a person accesses an HTTPS site, the computer and the Web server exchange and mathematically verify cryptographic keys to establish a secure link. Any data exchanged over that link is then encrypted and is practically impossible for an attacker to decrypt.
Cavium’s new chip can perform the necessary mathematical calculations much more quickly and efficiently than a general-purpose processor inside a Web server, making it cheaper to secure Web traffic, says Jeff Pangborn, the company’s principal engineer for networking hardware. “The people operating data centers are very concerned about efficiency and how much power they use,” he says.
When designing an embedded system choosing which tools to use often comes down to building a custom solution or buying off-the-shelf tools.