A decade ago, John Giacomoni was working as a professional research assistant in the Software Engineering Research Lab at the University of Colorado Boulder when the group took on a problem they couldn’t afford to solve.
Giacomoni was working in Professor Alexander L. Wolf’s lab and their task was to build a system that could secure the campus from electronic attacks. As the scope of the project expanded, they soon discovered the specialty hardware they needed to continue was a budget buster.
But necessity is the mother of invention, and the team had a need. Unable to buy the expensive hardware, Giacomoni became convinced he could build an affordable software solution for his costly hardware problem.
Convincing others would turn out to be the real problem. “Funding agencies and investors, except for a handful of Boulder angels, universally said, ‘No way; no how. You guys are crazy,’” Giacomoni remembered.
Ten years later, Giacomoni has proven he wasn’t crazy — quite the contrary. The affordable solution he envisioned as a CU-Boulder research assistant became a reality with the help of former CU-Boulder Assistant Professor Manish Vachharajani.
The pair went on to co-found a company based on the technology, LineRate Systems, which they sold earlier this year to F5 Networks. The acquisition price — nearly $125 million according to filings made with the Securities and Exchange Commission — was the largest exit of an networking infrastructure company in the Boulder area in years.
Doing away with the big metal box
Giacomoni’s “aha” moment was the discovery that, by writing carefully structured software programs, anyone could get ubiquitous general-purpose processors, which are relatively cheap, to deliver the same performance as software written for more expensive specialized network processors. Once LineRate was formed, Giacomoni and Vachharajani were able to show not only that the idea worked but that it could be a cost-effective solution for a variety of problems beyond the original impetus: foiling electronic attacks. Companies like LineRate helped drive the revolution in the field now known as software-defined networking.
In the world that existed before software-defined networking, companies with networking needs went out and bought a big, expensive “piece of metal” that contained a hardware-software intertwined package. But once companies outgrew capacity of that big metal box, their only option was to toss out the whole shebang, both software and hardware, and buy a new one.
“By decoupling the software logic from the underlying hardware and moving to a general-purpose platform you get two key benefits compared with buying boxes,” Giacomoni said. “In addition to getting centralized command and control over your data center, you can load the software on any server even cloud-based servers and you can upgrade the hardware and software independently unlike in the big-box model where you are forced to upgrade both at the same time.”
Trying to catch a break
It took Giacomoni about three years to prove his new vision of network processing could become a reality. By that time, April 2006, Giacomoni was a graduate student and he used his newly created software algorithm, called FastForward, in his thesis.
With his master’s in hand, Giacomoni pushed forward with his vision. He enrolled in a computer science doctoral program at CU-Boulder where he teamed up with Vachharajani, an assistant professor in the department at the time, who had experience with similar technologies from his own doctoral work at Princeton.
With Vachharajani serving as Giacomoni’s adviser, the pair sought research funding to prove that the FastFoward algorithm was correct and that it could be used for a wide range of networking needs. But traditional science funding agencies shied away from the project — and, later, so did investors — because they were reluctant to believe that it was even possible for general-purpose processors to do the job.
But the pair got two big breaks at the end of 2008. First, the Denver-based online photo-sharing website Photobucket was willing to confirm they needed what Giacomoni and Vachharajani were trying to create, which was critical for luring investment.
Then, CU’s Technology Transfer office awarded Giacomoni and Vachharajani a $50,000 proof of concept investment to incorporate, rent space, purchase vital equipment and cover a modest stipend so that Vachharajani could build a functional prototype while Giacomoni continued to pound the pavement looking for both customers and investors.
“The proof of concept grant was really just so critical,” said Kate Tallman, senior director of technology transfer for CU-Boulder. “They needed equipment to build and test their software. Without that equipment, they would never have been able to collect crucial data. They also needed independence to attract customers and investors”
Persistence of vision
The work finally began to pay off in late 2009, when LineRate got its first angel investor, a Boulder local, whose interest attracted other angel investors. Ultimately, they raised $575,000 in angel financing for the company, allowing Giacomoni and Vachharajani to hire seasoned networking developer Mark Terrel, create a production-ready product, and recruit experienced startup leader Steve Georgis to serve as chief executive officer, which in turn, attracted a second round of funding totaling $4.75 million.
By the time Giacomoni — who is now finalizing his Ph.D. dissertation for its final submission — and Vachharajani sold LineRate Systems, the company employed more than 20 people. F5 Networks, the company that acquired LineRate, was one of LineRate’s largest competitors.
“John and Manish were incredibly persistent in going after their vision,” CU-Boulder’s Tallman said. “They met with dozens of entrepreneurs and investors across the county who did not believe that their technology would be possible. Because they stuck with it, they were able to recruit this incredibly talented team of entrepreneurs and investors right here in our community who made their vision a reality.”