When a storage vendor goes shopping, privately held DataDirect Networks Inc. (DDN) often comes up as an acquisition...
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
target. DataDirect Networks, which specializes in high-performance computing (HPC) storage systems, has been considered a possible acquisition for its OEM partners Dell Inc., Hewlett-Packard (HP) Co. and IBM, and has even been linked to rival EMC Corp. We recently talked to DataDirect Networks' CEO Alex Bouzari about the vendor's plans for its business and technology.
SearchStorage.com: Where has all the acquisition speculation come from?
Alex Bouzari: That started very recently, I'm not sure what's driving it. We have absolutely no intention of getting acquired. I think it comes in part because we've been successful in addressing needs of scalable unstructured data environments, and there are more and more applications like that out there. We have the IBMs and Dells and HPs of the world reselling our products, and that in turn is getting people to think, 'Hey if these guys are doing so well and big companies are reselling their products, it means somebody must be thinking of buying them or they must be thinking of getting acquired.' But I really think it's the fact that our revenue is scaling and more of the big guys are coming to the conclusion that they can't do it themselves. They need DDN to get to these emerging markets where you have to store lots of data and there's a bunch of attributes around it that you need to have, and DDN is the only one who has these things.
Last year was a tough year for most storage companies. How was business for you in 2009?
Bouzari: We grew last year. We didn't grow as much as the prior year, and we didn't grow as much as we're growing this year. But we were profitable. It was a difficult market, and many companies shrank or lost money. We were able to weather some of those issues because the markets we serve aren't as impacted by the global economic challenges.
This year we're having a phenomenal year. The first half of the year has been the best year we've ever had by far, as measured by any metric. We're having a great revenue ramp in places like life sciences and video surveillance, for instance. These are markets that didn't really exist a few years ago, and here they are today with huge amounts of content and data that needs to be managed with small IT staffs. We're growing nicely, and the future looks bright.
In 2008 you talked about going public, but the IPO market disappeared. Are you thinking of an IPO again?
Bouzari: We got things ready, and in early 2008 the market started wavering and we didn't like that. We've always taken a long-term view of things and been customer-centric. Taking the company public in an environment where you have lots of volatility means the company's management team has to spend a lot of time with financial analysts and investors, and explain and do a lot of work that is not conducive to serving the needs of customers.
The company is now profitable, we're generating cash and have all the financial resources we need to expand our product, so taking it public really wouldn't add anything at this point and it would be a distraction. The markets today are probably as jittery and voluble as in early 2008. So we're in no hurry to do anything to take our focus from customers.
What's your product roadmap look like?
Bouzari: Partners started shipping our S2A6620 midrange system in the later part of last year. That product has lots of capabilities in terms of mixed I/O as opposed to just bandwidth as DataDirect has always been known for. We also started shipping our SFA10000 high-end product, which can handle up to 1 million IOPS. It gives customers the ability to run file systems and applications straight into the appliance.
Expanding our product line into the midrange is something that will give our customers the ability to benefit from all this technology we've been developing for the past 10 years, but until recently was only at the high end of the market. With the 6620 we're making it available for midrange customers.
The overriding theme on the technology front for us is we're moving from a bandwidth-only play to a mixed I/O play. More and more, we're seeing customers who need both — they need lots of bandwidth for content-rich applications, but they also need IOPS for their mainstream IT needs. And they would like to have one system that can deliver both. We have a fairly extensive roadmap that continues down the path of mixed I/O environments.
We also have a file system called WOS [Web Object Scaler], which is aimed at environments where you have a very large number of objects that you need to store, replicate and make available in a broad manner. We're starting to get that into the field, and that's something we will be developing on a go-forward basis.
When you talk about midrange, do you mean a scaled down HPC system or a mainstream midrange storage system?
Bouzari: We have a few hundred of these 6620s out already, and we're seeing people use them in deployments for mainstream enterprise applications and for content-based environments. It's for performance, but not extreme performance the way HPC or the big newsroom editing or post-production systems were doing it.
This is a system that will do 1 or 2 Gigabytes per second bandwidth, up to 100,000 IOPS; capacity-wise it will do up to 240 TB, but it can be much less. It's a midrange price point, not an entry-level $10,000 box. We spent tons of engineering resources taking the DDN operating system and all the algorithms we've developed to handle these complex content-based environments for high reliability, and it took us a couple of years to shrink it down into this midrange product.
How do your partners Dell, IBM and HP position DDN against their other midrange platforms?
Bouzari: We're not really seeing a whole lot of confusion in having them choose between DDN products and other products they carry. Typically, if the environment has some content in it, then it's easy for them to decide it will be the DDN product. If the environment is just pure traditional, small, random I/O traditional IP, then it's going to be one of their existing products. If it's both content and traditional IP, and there is a growth element in the environment where the customer is scaling, then it's DDN. If it's typical 'Well I'm going to grow 10% a year in capacity,' then it's not going to be DDN. So far it's been pretty clear when to sell DDN and when to sell their other products.
I would say the places where we're getting traction are in nearline kind of tier 3-/tier 4-type applications, the entertainment industry, life sciences, energy, video surveillance applications, universities … markets where people need a good price performance ratio but need it in a smaller size than hundreds of terabytes or tens of gigabytes per second.
In some cases we're seeing our partners position the 6620 into NAS environments by adding a third-party file system or their own file system on top of it. And they also sell it into environments with unpredictable growth; in other words, they're not sure how much bandwidth they'll need six months or a year from now. It gives them the entry level into DDN technology, and gives them the ability to grow from there.
Are people using the WOS product to build storage clouds?
Bouzari: That market is just starting. There's confusion in people's minds about what the cloud is and how to leverage it. The places we're getting traction with WOS is where people need to store and retrieve a large number of objects in a distributed manner. It could be a telco application where you have people trying to keep track of tens of millions of user IDs worldwide. It could be applications where you have hundreds of millions or billions of records that are constantly changing or being updated, and they need to keep track of that in a distributed manner. We have a single global namespace that can scale to hundreds of billions of files. It's globally distributed, it can do millions of file operations per second, so it's not really a notion of IOPS, it's more a notion of file operations. It basically does what an Amazon S3 can do, but it can scale orders of magnitude more and it's far more cost effective when you get a few tens of terabytes.
Do you need to add management software like snapshots, thin provisioning and so on, or is that not required for your markets?
Bouzari: Our migration to the SFA platform provides us with a path toward a robust enterprise feature set. The product set will evolve going forward to accommodate a broader class of applications, which will include optimizations for enterprise applications, while maintaining the core DDN design focus on performance and scalability.