Zmanda Strives to Crack Open Backup

While IBM-led Aperi brought attention and confusion to the role of open source in storage in the past year, smaller companies are also working to close the gap between open source and storage. (See Aperi Appears Amid Questions, Gang of Five Counters Aperi, and IBM: Aperi Lives.)

One of those is Zmanda, which markets a commercial version of open source Amanda (Advanced Maryland Automatic Network Disk Archiver) backup software, as well as services and support. (See Zmanda Protects Open Source and Zmanda Backs Up MySQL.) While Zmanda's products remain a far cry from what it will take to challenge enterprise storage backup from the players like EMC, CommVault, IBM, and Symantec, CEO Chander Kant says he has those companies in his crosshairs.

There are others dabbling in open source storage, such as file systems vendor Xinit Systems and content management firm Alfresco. (See Alfresco Adds Content Management.) But Zmanda provides the key storage application -- backup -- and its roadmap makes it "potentially highly disruptive in the enterprise and service provider markets," according to analyst Brad O'Neill of the Taneja Group.

Today, the only application Zmanda supports is MySQL, hardly the database of choice for most SAN administrators. But Kant says support for Oracle databases and Microsoft Exchange email is coming early next year, with more enterprise application support to follow.

"We're marrying open source and storage software," Kant says. "Our goal is to become the data protector for the open-source ecosystem, but that isn't to say we won't support Oracle or Exchange. We're working aggressively on open-source APIs that applications can plug into. If your environment is heavy with Linux applications, you can go to Zmanda today. If you're Oracle or Exchange heavy, you have to wait six months."Kant and his team have storage roots. He worked in business development at Veritas and was a storage product manager for SGI before starting a Linux services firm. Zmanda's VP of product marketing Dmitri Joukovski was a product manager for EMC's Legato NetWorker.

Kant got the idea for Zmanda while running LinuxCertified, which sells Linux products and services in the San Franciso area. "When my clients started looking for backup, we would say you can get Veritas, Legato, or we can come up and install an open source project called Amanda and sell you services," he says. "Every time they would take the third option."

Kant got $5 million in funding from VCs BlueRun Ventures and Canaan Partners to start Zmanda in early 2005. He hired a team of Amanda developers to make it more reliable and scalable. Since then they've added support for multiple tape drives, and Windows and Unix clients and servers.

Zmanda uses the open source pricing model. The software is free but you pay an annual subscription for service and support. Amanda's backup and recovery product cost $100 per backup server for basic and $250 per backup server for premium service levels, and Zmanda Recovery Manager for MySQL costs $250 per database for basic and $500 per database for premium support.

Zmanda's developers have a long way to go before matching the enterprise backup software for functionality. Kant's team is working on making its applications easier to install, manage, and configure, and must improve virtual tape library (VTL) support. Interoperability testing is also difficult and costly, especially for startups.And the marriage of open software and storage could be tough to arrange, considering the two aren't even dating.

"A lot of storage people don't get open source, but open source only refers to the development approach," analyst O'Neill says. "The code is shared and anyone can work against it. Zmanda can move into a broader range of applications; the challenge is finding partners who want a vested interest in controlling and leveraging data protection."

O'Neill says Zmanda's cross-platform support can bring it opportunities in archiving and managed services. "You order servers from Dell, blast Zmanda software across those servers, and you've just built yourself a nice open source, disk-based archive," he says.

Kant says if Aperi is successful, it can work in his favor. But it will be a long time before it has any impact.

"The current focus of Aperi is more towards storage switches and lower level devices," he says. "So it does not have a short term impact. But strategically we are marching towards a common goal. Aperi will enable us to talk to storage devices using open APIs."Dave Raffo, News Editor, Byte and Switch