Wednesday, September 17, 2008
Interview with Derek Gascon, Caringo
Austin-based Caringo (www.caringo.com) is a data storage startup which is applying content addressable storage (CAS) to commodity hardware. The firm's products are used by companies for storing archival and other data. Caringo is venture backed by Vodafone Ventures and Austin Ventures. To understand the firm better, we spoke with Derek Gascon, VP of Marketing for the company.
For those not familiar with Caringo, can you tell us how your product fits into the storage market?
Derek Gason: Caringo was started in 2005, by one of our founders, Paul Carpentier. Paul invented the technology called content addressable storage, at a company called FilePool. FilePool was acquired by EMC, and became Centra, one of the firm's most successful product lines in history. Paul had left FilePool prior to the acquisition, and had been a customer of Centera, and realized that there were inherent shortcomings in the original design of the product. He went back to the drawing board, rearchitected a solution, and the result is Caringo's CAstor. CAstor is build around content addressable technology, and is an advanced architecture over what was originally put into the market in 2002. Content addressable storage, if you're not familiar with it, is the storage of content such as unstructured data--file data such as word processing documents, spreadsheets, images, audio, video, that kind of content. It's anything that is not stored in a database. That data is stored in an object form in the storage environment. CAstor is essentially built o be a very robust, and scalable storage environment that can scale from 1 terabyte to the petabyte range, and do so very simply.
The very unique thing about CAstor, is that it has a software-only approach to storage. What I mean by that, is our software will actually run on any x86 architecture. That takes a lot of the challenge that customer have, in who their hardware vendor is going to be, and separates that from storage. You can select CAstor on the server vendor of your choice--whether that is Dell, IBM, HP, or a white box server, install CAstor on those servers, and build your storage infrastructure for unstructured data.
Who are your typical customers and how are they using your products?
Derek Gascon: It's interesting in that our typical customer really crosses a number of industry segments. The traditional content addressable segment of the market is really mostly focused on archiving content that is rarely going to be touched. What has been advanced with CAstor, over previous CAS technology, is that we're able to offer very high performance CAS. That allows active content to exist in the cluster, while also addressing the archive requirements around regualtory compliance and corporate governance. Because of that capability with a single product, we find that we have customers that are in the traditional archiving space--such as healthcare, for storing images and archiving for those as long as they do--and also those using CAstor as the infrastrucuture behind their software as a service offerings over the Internet. We have customer offering consumers online storage of rich media content, to distribute and share directly from a CAstor cluster. That's very different--other CAS technology wouldn't even be considered in that environment. Johns Hopkins is storing a lot of genomic research data onto CAstor, and reusing that data right away, even though they're also using CAstor for the long term storage. The original data generated there will have to be archived over time, even after it's initially been used for research on the front end.
There's lots of competition from EMC Centera in this space, also big OEMs like HDS, HP, and IBM-- how are you finding traction here as the small startup among the big guys?
Derek Gascon: The big guys certainly know about us. Paul, our founder, sold his initial technology to EMC, so they're aware of our capabilities. And, we have been seeing a lot more competition in active engagements between EMC and ourselves. We've beat out the Centera product in at least one customer account. But, it's not so much the big guys we're hitting all the time. We're also competing with companies in the clustered, NAS global namespace area like Isilon, Panasas, and IBRIX, where we get considered more often tha the traditional EMC, HDS, or even IBM technologies in those customer opportunities.
What drove you to using commodity hardware, rather than building an appliance or selling hardware?
Derek Gascon: One thing we noticed when we were building the technology, is that a customer is essentially buying storage and paying premium the way CAS is bundled. They are paying more for storage than they would pay for an individual disk drive. Even though we've seen the cost of drives continue to decrease, as density grows, customer aren't necessarily seeing that reduction in cost passed along to them from the big vendors. Those vendors are essentially bundling storage together with a controller array, and selling it at a premium. What we wanted to do was to give the cost advantage of the decrease in price of storage, so that users could realize those economic benefits. We've really stayed away from building an appliance, which is just building in overhead into the distribution channel, and into the organization's overhead.
What's your go to market strategy with the products?
Derek Gascon: Our primary objective in go-to-market is to build up our resellers. As a startup, we have taken the approach of going direct to build our customer base. But, the real objective is to get to the indirect model and lessen or focus on direct sales. We're looking at about 50/50 in terms of the breakdown of direct and indirect.
How big is the firm, I see you're here in Austin?
Derek Gascon: We have offices and our headquarters in Austin, and we also have offices in Belgium--where one of our founders is located--and San Diego, where our CEO is located.
Finally, what's your assessment right now of the storage market and what's next for you?
Derek Gascon: The storage market is actually doing very well. There's lots of awareness and visibility around affordable storage, to address the explosion in the unstructured data world. That recognition has brought us into higher consideration, as customers look for more affordable products. Going forward, customers are looking for not just affordability, but also the long term total cost of ownership. What clustered storage gives is a self-healing, self-managing infrastructure. You don't have to spend so much on management or administration time. You also see customer looking at the market in terms of how they're dealing with new data generation, and how to effectively manage and distribute that data. That's something CAS lends itself well to.