Five academics have developed an alternative Tor network (HORNET) that can handle traffic up to 93Gb a second, while maintaining your privacy.
The HORNET system is more robust than existing anonymity networks such as Tor to passive attacks and offers higher node speeds for an "almost unlimited" variety of sources.
It is the brainchild of researcher Chen Chen from his University Carnegie Mellon. The researcher developed HORNET together with Daniele Enrico Asoni, David Barrera, and Adrian Perrig of the Federal Institute of Technology Zurich, and Greek George Danezis of University College London.
"Unlike others applications Onion routing, HORNET routers do not maintain per-flow stages or run computationally expensive processes to forward data, allowing the system to run smoothly as new clients are added," the team wrote in the paper HORNET: High-speed Onion Routing at the Network Layer (PDF).
"HORNET offers payload protection by default, and can defend against some global espionage attacks.
"It is designed to be extremely efficient: instead of maintaining stages in each relay, connection status (such as onion decryption keys) performs internal header packet transfers, allowing intermediate nodes to run faster even on a large number of clients. ”
The group argues that HORNET has such a low latency state of anonymous protocols as Dovetail [PDF] without the “significant degraded security safeguards” that can eventually be bypassed by attackers such as government services which have global data collection capabilities.
The speed of HORNET is possible in part through the use of symmetric cryptography for the transmission of data that do not require per-flow stages at intermediate nodes.
"This design allows HORNET nodes to process anonymous traffic at 93 Gb per second, adding a minimal charge per additional processing of anonymous channels," the team said.
They also state that the platform opens the way for anonymous Internet through small compromises on the size of the headers, something that will greatly benefit security and maintain high performance.
