Caringo Uses USB To Build Secure CAS Storage Appliance
That start-up, Austin, Texas-based Caringo, is planning to unveil its first product, CAStor, which allows solution providers and ISVs to turn any industry-standard server into a content-addressable storage node.
Content-addressable storage, or CAS, is a technology which turns a fixed-content data file, which is data that once stored does not need to be changed, into an object using a cryptographic algorithm to develop a unique object based on the content of the file to be stored. Because each object is unique, even a slight change to the content of the file changes the hashed object, making it possible to eliminate duplicate files. Using CAS technology also prevents the unauthorized reading of stored information, and allows administrators to see if data has been changed from when it was originally stored.
The current leader in the CAS space, EMC, got its start in 2001 with the acquisition of a Belgium software company, FilePool. EMC used that company's software to develop its Centera appliance, which allows data to be stored using CAS technology for compliance purposes.
One of the founders of Caringo, CTO Paul Carpentier, was also a founder of FilePool along with Jan Van Riel, who is now EMC's director of technology for the Centera line.
Mark Goros, CEO of Caringo, said his company is looking to bring a disruptive CAS technology to market in that the company's CAStor is a software-only product that works with any x86-based server, unlike other CAS products that are sold as pre-packaged CAS appliances, such as the Centera.
CAStor is also suitable to work with fixed-content files of any size, said Goros. This is an important differentiator from appliances such as Centera, which has trouble working with multiple small files such as check images, he said.
The operating system and the software for CAStor is sold pre-loaded on a USB memory stick, Goros said. To use it, the solution provider plugs the device into the server's USB port, and 60 seconds later they have a CAS node. This process can be repeated for thousands of servers, all of which can be connected together in a CAS cluster, he said.
Once so configured, each data file is stored at least twice in the cluster. For disaster recovery purposes, users can specify that additional copies are stored at a remote location. "If there is a disaster, they can bring the remote site on-line and then run their applications," he said.
Another capability unique to the CAStor is an upgradeable hashing algorithm, said Carpentier. Increased computing power means that over time the cryptographic algorithms used to hash an object from the contents of a data file are more and more susceptible to being broken and making it possible to modify or delete the data. "Over the long term, the only way to ensure the integrity of the data is to upgrade the hashing technology," he said.
Should customers need to upgrade or replace a CAStor node, they can tell the server to retire itself, at which time it stops accepting content, moves its existing content to the network, and then goes out, said Carpentier. Or users can just yank the server out, and the other servers will rebuild the data, he said.
The company is looking to ISVs, system integrators and custom system builders as the main go-to-market route for CAStor, which is expected to start shipping in late June, said Goros.
"Initially, we may sell direct to end users to try the product out and see what new applications people will come out with, like medical imaging, video storage, surveillance and so on," he said. "We believe the CAS market is only just beginning to take off."
So does Don McNaughton, sales manager at HorizonTek, a Huntington, N.Y.-based solution provider.
"Customers are asking about CAS a lot in conjunction with other things like e-mail compliance," said McNaughton. "These are typically larger deals with multiple pieces, including data encryption. Customers are buying a lot of hardware for such deals, so being able to use CAS to cut data storage requirements is important."
CAStor is list priced at about $500 for each hard drive that is added to the cluster. That averages to about $2 per Gbytes of storage, including the cost of the hard drive, when used with 500-Gbyte SATA drives, said Goros. That does not include the cost of the x86 server, he said.
Two dollars per Gbyte seems to be a very low price when compared to EMC Centera and other products, said McNaughton. "And end users like that they can use software on any generic hardware box to get a solution," he said.
Caringo joins a market that is quickly getting filled with vendors looking hungrily at the need of customers to find ways to safely store data to remain compliant with regulations that require that data cannot be modified or deleted for a certain time period.
Woodland Hills, Calif.-based Nexsan Technologies, for instance, last year introduced its Assureon storage appliance for use in compliance and information lifecycle management (ILM) applications.
IBM recently beefed up the compliance capabilities of its DR550 line of appliances, while Hewlett-Packard has added that capability with its Reference Information Storage System (RISS) appliance line.