Huai discussed plans for supporting Microsoft and virtual server integration, as well as how the data deduplication virtual tape library (VTL) business is going now that major OEM partners IBM and EMC have declined to use FalconStor's Single Instance Repository (SIR).
What's FalconStor's strategy for cloud storage?
ReiJane Huai: The cloud should be viewed from two vantage points – scalability and distance. A cloud can be overhead and around the globe. That's how scalable the IT cloud should be. The first question service providers have to ask any vendor is, can the product go long distance, can it go around the globe? Does it have the ability to manage petabytes of data? Do you have customers storing petabytes today? You have to be able to prove the scalability of the infrastructure because the collapse of one item in the data center has an impact on many, many customers. It has to be very, very, very reliable. You have to have something like the Microsoft stretched cluster, designed with that purpose, to achieve recovery objectives and scalable distance.
What's the biggest problem to be solved in data protection today?
Huai: The biggest problem isn't the ability to protect data. The biggest problem is the reverse, recovering it. There are too many people touting the ability to deliver data to another location or push it into the cloud. Very few people guarantee how accurately and how quickly data can be reproduced. The way we position our solution is in a very balanced manner. Not only do we show how data can be protected -- that's an absolute given -- but we also have a restore warranty. For example, with CDP [continuous data protection] you can recover, whether it's email or a database, within 10 minutes, no hedging. When it comes to disaster recovery, we guarantee 30 minutes.
What happens if someone can't recover?
Huai: I haven't seen or heard of customers having difficulty. It's because the method we use is fundamentally different. The CDP server pushes data to two locations. Later, if you need to retrieve information from Exchange, you don't need to do a data transfer from one place to another -- you can access your data on the CDP server. So we'll guarantee the ability to recover with that in 10 minutes, or show us the door.
So you'll give money back?
Huai: No problem. But so far we haven't had it happen, because we use a totally different method.
This also relates to our NSS [Network Storage Server] solution and how we're planning to support Microsoft stretched clusters. This is a case about availability. If you have downtime, it's a disruption that will cost the top line and bottom line. The Microsoft stretched cluster can be loosely coupled , one in New York and one in downtown Boston, for example. Our NSS server will act as a vehicle between one of the cluster nodes' local disks and the other. We will be able to replicate the information from one side to the other, only transfer the delta changes, and make sure both sides are synced.
NSS will also have the ability to detect a disruption and fail over to the other site. Meanwhile, the server in Boston comes up automatically, and users don't see the difference. No recovery time. In our case the distance recovery guarantee is usually 30 minutes because when the primary site goes down, you have to bring up the disaster recovery site and within 30 minutes be able to provision all LUNs to the host. This is similar to what we do for VMware. With Microsoft Windows Server 2008 Failover Clustering deployed in a multi-site configuration we are able to provide automated application failover and failback on physical clustered nodes and Hyper-V based virtual machines.
Switching gears for a moment, how many customers do you have for your data deduplication, Single Instance Repository?
Huai: Hundreds of customers are running VTL with SIR.
Why do you think IBM and EMC chose to resell your VTL without SIR, while working with other companies to offer data deduplication?
Huai: Everyone has their own process when it comes to differentiation, whether it's partnering or owning a product. We certainly respect the decision made by any partner. I would point out that IBM and EMC still need to lead with our VTL as the interface between storage people and the backup system. As far as dedupe, it is a component that they can add on, and we're completely open to that. Also, Sun Microsystems has made the decision to go for our VTL along with our deduplication, and SpectraLogic is doing the same.
Has the changing relationship with IBM and EMC impacted your revenues in VTL?
Huai: We've seen no impact at all on VTL revenue. Our VirtualTape Library is delivered by EMC together with Quantum's dedupe.
What's your outlook on those OEM deals with IBM and EMC?
Huai: I'm not really in a position to make a directional statement on their behalf. I'm confident about our position in the industry, and we are working with many partners for our technology. We also have a reseller channel as a vehicle for distribution. We're not a single product company, and we're not a single partner company either.
It seems like you're putting more focus on NSS and CDP in recent announcements. Is that in response to what's happened with SIR?
Huai: No, that's just a perception. We have three product lines – VTL, CDP, NSS and another major line pretty soon. It just depends on the occasion when we make certain things available, like for VMworld our focus was on integration with NSS. You will see the announcement of other products in the very near future. You'll see interesting announcements about VTL, dedupe and how we're extending the SIR portfolio to support not only VTL integrations but others, as well.
HP's acquisition of LeftHand will put storage software running on commodity hardware even more into the spotlight. How does FalconStor plan to stay competitive there?
Huai: We certainly don't feel competitive pressure at this point, but we're very encouraged by HP's move. Because our virtualization is technology-agnostic, we will participate in the iSCSI area, as well as Fibre Channel and InfiniBand. We already have a presence in China because of our partnership with 3Com's subsidiary H3C. They OEM our technology and became the No. 1 iSCSI storage provider in China. Hopefully, with the right partners and channel, we'll see the same success everywhere else in the world.
With all the mergers and acquisitions in the storage market in recent years, have you been close to acquiring or being acquired?
Huai: This is a question I can only give you a political answer to because we're a public company. Over the last few years, because we have good cash flow, we've been doing stock buybacks – it's like investing in our own company. But this isn't to say we're not interested in acquisitions, but it has to be a good match, like a marriage.