Mapping the DNA of network data


7 Feb 2003

It is widely accepted that at Stanford University, neighbouring Silicon Valley in California and the birthplace of internet industry giants like Sun Microsystems, Yahoo! and Cisco, some 60pc of the academic staff have commercial interests in business and tech start-ups.

When I put this to Amit Singh, founder of one of the fastest growing IT networking companies in the world, he laughs and suggests it is closer to 75pc. This contrasts starkly with the Irish situation where the rate of technology transfer of an idea or project into a world-class business is perilously low.

Technology transfer offices in Irish universities in most cases employ just one person. Singh is speechless at this statistic. Can you blame him?

Singh was researching a PhD in bioinformatics at Stanford University when he began to wonder if the ability of computers to spot and compare patterns in DNA could be used to speed up computer networks.

The result is a little black box that reduces bandwidth use on corporate networks by up to 90pc and a business that goes against the grain in a time when start-ups have limited opportunities. “Basically we have a technology that can make a corporate network that is supposed to run at 1Mb per second run at 4Mb per second,” Singh explains.

Singh’s work at Stanford involved the use of computers to understand molecular biology, similar work to the mapping of the human genome.

He realised that the techniques used to find patterns in DNA sequences could be used to identify repetitions in continuous streams of network data and help reduce the amount of data being sent over networks to make things move faster. Initial testing resulted in data reduction rates of between 80pc and 90pc.

The technology transfer-friendly aspect of Stanford resulted in Singh getting access to venture capitalists and in 2001, the company received some US$10.5m in first round finance from Accel Partners and Foundation Capital.

The concept took hold and aided and abetted by industry veterans from Intel, Hewlett-Packard, Sun, Microsoft and Oracle, the company’s first product, the SR-50, began to spark the imaginations of chief technology officers (CTO) and IT managers across the US.

A further round of funding of US$20.5m from Mayfield Venture Capital in February 2002 brought Peribit’s capitalisation to US$31m, which Singh reckons should carry the company through the next four to six quarters to profitability, where an initial public offering is still envisaged.

The SR-50 is based on Singh’s patent-pending molecular sequence reduction theory, which employs a unique sequence of algorithms that identifies data flow and pattern similarities and eliminates repetitive data running over a wide area network (WAN). Singh believes that Peribit’s technology is striking a chord amongst CTOs and IT managers at a time when IT budgets are under considerable pressure.

Peribit established a local sales office in Ireland in the middle of last year under former 3Com country sales manager, Shane Buckley.

In recent weeks the company consolidated its presence in the Irish market through striking a partnership with local networking firm Lan Communications. Under the agreement, Lan will resell Peribit’s sequence-reduction products which can increase network throughput up to 10 times over existing WAN transmission facilities.

According to Buckley each SR-50 can be installed within 10 minutes by an IT manager to facilitate data and voice traffic and reduce network clogging.

At present, after only two years in existence, Peribit, which employs some 62 people in Santa Clara, is active with resellers spread throughout the US, Canada and Latin America and plans are afoot to enter the Asia-Pacific market.

Speaking about the means by which his DNA-based network sequencer speeds up networks, Singh explained that a typical WAN running at 128Kb per second could be fine-tuned to becoming a 2Mb per second network.

“We are also focusing on developing solutions for SMEs [small to medium-sized enterprises] and small home offices so you could say we mean to cover every extreme, from the largest corporate down to the smallest office.

“We are focusing on growing Peribit quarter on quarter and believe that within four to six quarters, we will be a profitable company. In our particular field, we were the only company to have increased our market valuation in the past year,” Singh notes.

The secret of his success? “We are not a threat to existing data carriers like 3Com or Nortel Networks,” he comments. “We simply co-exist insofar as we make it easier for their network to run. In fact, Nortel is a customer of ours. We prefer to be referred to as an alternative deployment, a means of optimising an existing IT investment.”

In what can be regarded as a lesson for Irish campus companies, Singh is adamant that technology research and development were key to capturing new markets. “We have a breakthrough technology, but that’s not to stop someone developing an alternative competing solution. We aim to stick to our knitting as far as R&D [research and development] goes. Technology development is a major focus of our work in adding more and more value to the enterprise networking market.”

Pictured: Shane Buckley, president EMEA