XOsoft: Before and after Akamai

There have been two significant periods in XOsoft’s life. The first was before everyone knew Akamai and the second, after everyone got to know who Akamai is. In short, everyone heard about the most successful issue in Israel’s history for a company that in zero time achieved a company value of $24 billion (now reduced to an eighth of this sum).

XOsoft founder and CTO Leonid Shtilman says in the period before Akamai, not one investor wanted to talk to him and he had great difficulty finding financing for the company. “With all due respect to Israel’s high-tech industry, in many respects the attitude is provincial. If a similar company does not exist in the US, they have difficulty investing in you,” he says.

Things changed in May 1999. Even before Akamai’s issue, people started talking about the company as the next buzz word and doors suddenly started opening to Shtilman. The $2 million seed money was quickly provided by Neuron Venture Capital.

Shtilman, a mechanical engineer by training, became a professor of computer science at MIT, specializing in mathematics. He also served as a NASA scientist, set up Mutek Solutions in 1996 and was a director of Constellation 3D. Shtilman founded XOsoft together with vice president, R&D Dr. Michael Zakharov, a former CTO of the Russian Space Science Internet network, Dima Barboy, Leonid Kogan and vice president, engineering Dr. Sergei Kaplan, former CTO at NetReality.

The idea for setting up XOsoft was raised in November 1998. At the time, the company considered entering into management of large networks, but the founders noticed the many problems in Internet “collisions” in Internet information. Shtilman consulted with his cousin, Geo Interactive general manager Eli Reichman, whom he had raised from a small boy. “This kid I raised helped me understand the problems that are created on the Internet,” Shtilman says.

ISPs as customers

Since 1998, at least 12 competitors have sprouted up, but Shtilman saves his spears for Akamai. Incidentally, he says that he had not heard of Akamai when XOsoft was set up and development proceeded independently. Shtilman heard about Akamai in a conversation with Yossi Vardi in February 1999. “I was greatly shocked to learn there was another company with a similar idea to ours,” he says.

Despite the similarity between the companies, XOsoft’s solutions differ from Akamai’s. Both companies’ ideas are very simple, but the solution is anything but simple. All content delivery companies base their technology on a simple fact: surfers using sites with servers that are physically closer achieve higher speed transmission. Akamai does this by deploying 7,000 servers throughout the world (in 800 sites), while XOsoft accomplishes it through synchronizing technology for dynamic information.

This is where it becomes more complicated. Synchonizing static sites is not very complicated but it is an entirely different matter when it comes to dynamic content, since it does not exist as a file and cannot be transferred from one place to another. For example, when using a personal Yahoo! site (My Yahoo!), the information becomes active and is not based on files that can be spread over servers. Shtilman says that Akamai is capable of calculating where the surfer’s nearest server is located, but its solution acts only on static content, without synchronizing dynamic content.

Shtilman had an idea, using the saying “It should be possible to synchronize between computers”. The key problem in making synchronization effective is knowing how to synchronize only the changing information, without wasting expensive time and bandwidth synchronizing the entire file each time. Shtilman say there are other problems, such knowing when a file has been changed, but he is confident he can solve them.

Shtilman: “Akamai’s solution, in effect, stalls development of the Internet. Site designers currently need to design sites that produce better performance and do not build them very sophisticatedly. They keep them simple for easier network transmission. Not many sites have flash, since it is dynamic.”

XOsoft made a strategic decision not to conduct a head-on marketing fight with Akamai, but rather through infrastructure owners. XOsoft products are not available for direct sale to site owners, but through ISPs, who obviously prefer technology which will bring in more traffic. Shtilman looks ahead to the cellular market and says that his technology is the best for this market, characterized mainly by dynamic content and that it is suitable for a wide range of formats.

Akamai’s average customer, a site manager, pays the company almost $120,000 a year for its solution. Shtilman is interested in the sum being paid by the site to the ISP, which will earn almost $10 million a year, with a small sum to XOsoft.

Shtilman is not afraid of taking customers from Akamai, since, he claims, he hears everwhere that Akamai is a sort of “parasite” vis-a-vis ISPs, which maintain the infrastructures, while the company takes their customers.

Programmers should earn more than CEOs

In its last financing round, XOsoft raised $17 million at a company value of $47 million. Seed Capital Partners, Goldman Sachs, Draper Fisher Jurvetson Gotham (DFJ Gotham) and Comsor were among the investors.

Shtilman wants to double the 68-strong staff by the end of this year. To that end, the company is raising $30-40 million at a company value of $100 million, after money. However, Shtilman says that not every programmer is suitable. “I’m not looking for someone who wrote programs in the army for three years. I might give him a job, but it’s not what I’m looking for.

”I’m willing to pay a much higher salary than the accepted rate for a CEO immediately for a low-level Unix programmer. I’ve been looking for such people for six months and have come to the conclusion that they simply don’t exist in Israel. This is actually the reason we set up a center in California.”

Shtilman has an exceptional attitude in Israel towards promoting his programmers. He believes that an experienced programmer should earn more than his manager, like in the US. He claims the Israeli promotion route is problematic in that after writing codes for three years, a programmer has to climb the managerial ladder in order to improve his salary, which naturally adversely affects the quality of programmers in the country. “I’m willing to take a manager from another company to work for me as a programmer and pay him considerably more,” he says.

Published by Israel's Business Arena on 5 February, 2001