Taking data retrieval by storm

InfoCyclone’s technology protects sites from crashing when the system is overloaded.If surfers think they are tired of waiting several seconds for answers to come back from Web sites to queries they have sent over the Internet, imagine how tired the …

InfoCyclone’s technology protects sites from crashing when the system is overloaded.If surfers think they are tired of waiting several seconds for answers to come back from Web sites to queries they have sent over the Internet, imagine how tired the giant content providers – such as CNN in the news category, Amazon in the retail business or eBay the auction king – are of paying the tens of millions of dollars it costs to lease and maintain the hundreds of servers where the data are stored and from which they must be retrieved on demand.



“Picture a computer server as a huge chest of drawers, each one containing information that is kept until someone wants to take it out,” said Ran Giladi, co-founder and CEO of InfoCyclone, a Tel Aviv-based start-up. “Next, compare each online query to a command to look for and open one specific drawer, and send the contents back over the Internet to the impatient user.”



Giladi is far from finished describing the cumbersome process.



“A popular or busy global Web site – like Bank of America, CNN or Yahoo!, for example – must pay for entire server farms just to store the vast amounts of data they compile, and on top of that even greater sums to employees who maintain the sites: the people who make sure the software directs the queries to the right ‘drawer’ and the correct answer is sent back, as quickly as possible.”



Facilely switching similes, Giladi likens his technological solution not to furniture but to a meteorological phenomenon. “Now imagine that the data need not sit still inside virtual drawers, but can swirl around above the ‘dresser,’ ” he said, like motes of dust made visible in a room by rays of sunlight coming through the window.



The data would, of course, be more organized than that: a better metaphor might be a mini-tornado, in which the data revolves in the funnel in a manner that is fixed and not haphazard, thus enabling “reaching out and grabbing” the bit of desired information the essence of the retrieval process.



According to Giladi, “InfoCyclone’s technology will enable Web sites to execute hundreds of thousands of queries per second – which is 100 times more queries per second than currently possible. This is not just a matter of speed, but it is also the optimal solution for load spikes, protecting sites from crashing when the system is overloaded.”



Giladi recalls what happened on the eve of the 2000 U.S. elections. “CNN has the resources to handle thousands of queries per second, covering just about any and all instances of breaking news, catastrophes or crises anywhere in the world. But on election night, CNN could not keep up, and its site crashed. This is what InfoCyclone can prevent.”



Giladi reeled off a list of businesses and organizations whose Web presences need the kind of application driven storage technology that InfoCyclone has perfected.



“The Web-based systems requiring support for high volume query transactions per second (100,000 queries per second and above) include search engines, e-commerce sites, large banks, e-trading brokers, news organizations and telecommunications providers, as well as emerging and future enterprises that will require either extremely fast query responses or the handling of a very high volume of query-dependent traffic.



“A prime example of a market that is sure to grow is Internet banking,” Giladi observed. “It is projected that online banking will grow 400 percent in the next few years. Yet already, Bank of America alone has 3,000,000 million online customers.”



While Giladi is naturally reticent about going into details that might reveal any intellectual property secrets, “the key to InfoCyclone’s core technology is our high volume retrieval engine,” he explained.



“InfoCyclone’s retrieval engine is an appliance based on state-of-the-art technologies such as retrieval processors, multi-gigabit per second data transfer channels, parallel distributed processing, pipelining, and content associative memory technologies. “The goal InfoCyclone has set for itself is always to be 10 times faster than the competition, and 30 times more worthwhile in cost efficiency.”



Because of the need for this kind and degree of reliability, Giladi anticipates another full year of testing before the company begins to sell its product and realize any revenue. But InfoCyclone has the full backing and confidence of investors that will see the company the distance.



Indeed, the team comprising and recruited by Giladi and InfoCyclone cofounder Michael Shurman is impressive. Giladi and Shurman were among the chief technologists behind Fibronics, a company that was instrumental in the development of Israel’s data communications infrastructure.



Subsequently, Giladi went on to become chairman of the communications engineering department at Ben Gurion University, the first and only faculty of its kind in the country.



“The BGU communications engineering program represents a tremendous pool of talent for InfoCyclone,” Giladi said. “By the end of this year, InfoCyclone will have 25 employees, the best we can get from all parts of the country. In fact, the reason we are located in the heart of Tel Aviv is because our employees commute by train from Haifa in the north and Beersheva in the south, meeting here in the middle.”