Features - Financial Plays Of The Week:

CEREVA FOLDS - FAILS TO GET STORAGE ARRAYS TO MARKET

High-end storage startup Cereva Networks decided to close its doors, a move that won't make it any easier for the clutch of competitive high-end startups such as YottaYotta, Zambeel and 3PAR to establish themselves. The company ceased its business operations, and is now in the process of settling and closing its business with the assistance of a small number of employees.

There are conflicting reports as to the reasons behind Cereva's closure. But the general consensus is that CEO Mahesh Ganmukhi declined to take further funding from the VC backers due to a lack of confidence in the product or its market prospects.High-end storage is a difficult business for small companies. Incumbent vendors such as EMC, IBM and Hitachi have spent years developing their architectures and getting them to the necessary relaibility levels. Cereva is based in Marlborough, Massachusetts, and had about 150 staff at the time of closure. It soaked up $137m in funding and a further $20m in credit during nearly four years of existence.

There's now one less competitor among high-end startups. The remaining players include Yotta Yotta, Zambeel, 3PAR and Spinnaker Networks. This failure, in a 'hot' sector of the market, is a warning sign to customers and to the industry itself. The investments required to establish new hardware architectures are sizable, and the time it takes to build and test them is lengthy.

Those four years may have been part of the problem. When the company was first incorporated in 1998 under the name Rajmataj, the industry was a very different place. And while the core product design was being worked during the following year, all the talk was about a market for giant storage arrays that service providers could use to consolidate their operations, in order to pass along cost-effective online storage services to their customers. But customers proved unwilling to hand over control of their data to the SSPs, and that market failed to take off.

The first real sign of trouble at Cereva came with the resignation of CEO Alan Lutz (from Compaq and Newbridge Networks) in May 2001. Lutz had joined early in 2000, and boasted that he would take the company's headcount to well beyond 400 within a year. Ganmukhi, a startup specialist who sold both his two previous companies to Lucent (they were switch maker Cascade Communications and optical network company Ignatus Communications), started to rationalize resources, and in September 2001 closed the $51M second round of funding. That same month, head of sales Timothy Lieto and chief financial officer resigned from the company.

The Cereva 5000, a switch-based storage array said to be capable of scaling from 200 terabytes up to one petabyte, did make it into the hands of a few users for evaluation purposes, but no firm plans for a commercial release were ever finalized. Meanwhile, potential competition in the shape of other startups, such as Yotta Yotta, 3Par and Zambeel, began to emerge. These are now starting to reach the market.

Cereva's investors include Matrix Partners, North Bridge Venture Partners, Oak Investment Partners, Sumitomo, Intel, Goldman Sachs and Worldview Technology Partners. Comdisco Ventures and Global Crossing Ventures, whose parent companies are both struggling for survival, participated in the first round but not the second. None are likely to see any of the $160M outlay again.

While the elimination of a competitor might help the remaining high-end startups, they're actually more likely to be hit by the negative implications that small companies aren't credible suppliers of high-end equipment. 3PAR counters this by pointing to its trio of strategic investors Oracle, Sun and Veritas that have certified that 3PAR's technology works with their own, and tested other potential showstoppers, such as migration issues.

David Scott, CEO of 3PAR, believes that Cereva was hurt by a constantly changing focus during the development process. "They started off aiming at content caching, then moved to service providers, then government users so we didn't come across them," says Scott. "We targeted enterprise customers from the start." 3PAR's beta testers include Prudential Financial, AIG and Oracle.

The depressed economy is making new startups reassess their target markets more carefully. StoneFly Networks originally intended to release a high-end storage switch, but repositioned it as a low-end device before the company was launched, due to market conditions. And Zambeel, while keeping its focus on the high-end, is aiming its first products only at more tolerant scientific and technical users, and delaying a broader push to sell to enterprise customers in the commercial mainstream. Zambeel's NAS-oriented product is now generally available. The University of Wisconsin was one of its beta testers.

Will Cereva's technology be acquired? It's unlikely that incumbents EMC, Hitachi or IBM will step in. Hitachi already has its own switched architecture in place, and EMC is working on its own migration strategy, but isn't likely to change the existing shared-bus architecture of Symmetrix any more quickly though Cereva's technology. Sun Microsystems is a possible acquirer, however, in the past it has bought small hardware firms and cherry-picked the technologies it wants from them Thinking Machines and Encore being two examples. As for the other startups, they might be interested in engineering expertise, but they are already too far down their own development paths to gain advantage from buying anything in besides which they are all concentrating on preserving as much cash as possible.

Courtesy of the451.com .