Azul Sytems Clarus Systems Go Networks Matisse Networks OATSystems OpTier Ping Identity Revivio Univa Virtual Iron

Azul Systems

By Jennifer Mears
Network World, 04/25/05

Location: Mountain View, Calif.

What does the company offer? The Azul Compute Appliance, which runs on custom-designed silicon that includes 24 cores on a chip, and virtual machine proxy software for application servers. The combo offloads demanding Java processing workloads from servers.

How did the company get its start? Former Nortel executives Gil Tene and Shyam Pillalamarri, and Scott Sellers, an entrepreneur and former Silicon Graphics engineer, founded Azul (originally named Chestnut Systems) in April 2002 to address performance bottlenecks that frequently are associated with Java-based applications.

How did the company get its name? "Azul" is Spanish for blue, and CEO Stephen DeWitt has a passion for bright blue hues because he believes they emanate good feelings. When he talks about Azul, DeWitt uses a Big Blue/Little Blue analogy, positioning Azul as a company destined for something big.

How much funding does the company have? An undisclosed amount of funding from a variety of investors, including Accel Partners, Austin Ventures, ComVentures, Red Point Ventures and Worldview Technology Partners.

Who's leading the company? Stephen DeWitt, who has a long history in the systems, networking and consumer software industries. Most recently, DeWitt was vice president and general manager of content delivery and edge computing at Sun, which he joined in 2001 with Sun's $2 billion acquisition of the bluish-named Cobalt Networks, the server appliance vendor he headed.

Who's using the product? Pegasus Solutions, Everett Consulting and EDS are among a handful of customers beta-testing the product. The appliance began shipping April 18.

Why is this company worth watching? Web services and service-oriented architectures are creating big headaches in data centers as IT managers struggle to figure out just how much hardware is needed to support these spiky applications. Typically, they've had little choice but to over-provision to make sure the infrastructure can support peak demand. And this has resulted in costly management expenses and underutilized resources.