The scientists from the University of California, San Diego and Purdue University say their Lossy Difference Aggregator offers the possibility of diagnosing fine-grained delays and packet loss as infrequent as one in a million at every router within a data centre network.

The system could be rolled out in today's router designs with almost no cost in terms of router hardware and with no performance penalty, claim the team.

The researchers say their technology could prove extremely important for algorithmic trading, where reacting to information on stocks microseconds before the competition can be worth millions of dollars.

"Every investment banking firm knows the importance of microsecond network delays. Because routers today aren't capable of tracking delays through them at microsecond time scales, exchanges such as the London Stock Exchange use specially crafted external boxes to track delays at various key points in the data centre network," says Alex Snoeren, computer science professor, UC San Diego Jacobs School of Engineering.

However, these external systems are generally too large and expensive to be added to every router in a data centre network running an automated stock trading system. The scientists say they have now come up with a cheap alternative.

In a paper presented at networking conference SIGCOMM, the team offer simulations and proof-of-concept code for measuring latencies down to tens of microseconds and losses that occur once every million packets. They are now looking into building a hardware implementation.

George Varghese, computer science professor, UC San Diego Jacobs School of Engineering, says: "This is stuff the big traders will be interested in. But more importantly, the router vendors for whom such trading markets are an important vertical."

You can read the paper here.