Testing the Cloud: SOASTA CEO Tom Lounibos (Part 2)

Sramana: What did you do after Knowledgeware?

Tom Lounibos: I went into small startup in the object oriented programming space called Digitalk. Our product was Smalltalk. When Steve Jobs went to Xerox park to find the mouse he found three things that day. One was the UI, one was the mouse and one was Smalltalk. We took an instantiation of Smalltalk our of Xerox park and we eventually merged our company with ParcPlace.

After Digitalk I met a gentleman named Ken Gardner and we started a company called Sagent Technology in the business intelligence or big data analytics space. Our competitors in the 90’s were folks like Business Objects. We built what we considered to be a better mousetrap. We also incorporated an ETL into our analytics engine that allowed us to pull data from a lot of different sources. It was like Informatica and Microstrategies combined as a single product. WE grew that to a 100 million dollar company and went public in the spring of 1999. I jokingly say that even the donut companies could have gone public if they had a billion dollar market cap. In fact Krispy Kreme went public the same day we did.

Sramana: I remember Sagent in the late 1990’s. What happened to that company?

Tom Lounibos: The company continued to grow and then missed a quarter or two before it was merged with Group One Software. I left in late 1999 to be an entrepreneur in residence at Crosspoint Venture Partners. A lot of our startups had been founded by Crosspoint Ventures. I was there looking at a lot of different business plans when I found a group of guys who had an idea but had no money. We started a company called Dorado.

Dorado Corporation was one of the first SaaS based companies ever developed. In 2001 we called them vertical service providers. Dorado was a loan origination system. The idea was to automate the mortgage process for banks. We supplied banks with an application and we were a black label solution for 10 of the top mortgage lenders.

At the same time Ken went off and build the next generation OLAP engine. The concept of real time is the business deliverable but you do that with an in-memory OLAP engine. Ken’s company developed the first one. I was on the board and Ken was trying to find applications for in-memory OLAP engines and he could not find any. I had an interesting dinner with Chase Manhattan in 2002 where they were ready to sign a very large 3 year contract with Dorado. They called me back to New York to help answer some questions about SaaS and our company.

When I was there the first question I received was “What happens if mortgage rates drop tonight and tomorrow morning 100,000 Chase customers rush to our website to lock in their rates?” It was essentially a scale question, ‘can a small company like you handle a big company like us?’. I froze up a bit and I had to say that in every test case there is a software, hardware and people component. This problem had all of them.

First, the current test tools of the day were not interactive. I flippantly stated that the multi-million dollar deal that Chase was giving us was not big enough for me to go ask Mercury to give us a 100,000 user test license. I knew that to simulate 4,000 virtual users would cost me a million dollars. I told him that in order to generate that load I needed hardware which could turn out to be 800 servers in a lab somewhere.

My point was nobody had a lab to simulate 100,000 web users. The whole idea of doing a test of 100,000 users would be to identify latency and application failures. There was no analytic engine that could pull all of the various data points together for testing. The moment I explained all of that to Chase I thought of Ken and his need for a real time OLAP engine application.

In many ways SOASTA was founded that day, at least in theory. We eventually sold Dorado and Ken sold his company to Oracle. Around 2006 we started looking at cloud computing and we realized that was the missing piece. If we could get our hands on thousands of servers in an affordable manner, in the case of Amazon around 40 cents per server, then we could spin those servers up, test a site, and then use our OLAP engine to analyze the results.

At that point and time there were 10,000 web and mobile apps out there. Today there are millions of them. Eventually everybody started asking how to prove their systems and ensure they work the right way. Performance has become very hot right now.

This segment is part 2 in the series : Testing the Cloud: SOASTA CEO Tom Lounibos
1 2 3 4 5 6 7