Sramana Mitra: Let me get a few things straight because I think we’re dealing with a fairly complex ecosystem here. I want to understand the real positioning of your company in that fairly complex ecosystem. We have the data center vendors, enterprise, and customers who run their own large-scale data centers. You are enabling these enterprise customers to layer on software around data center on top of the basic infrastructure provided by the data center vendors?
Momchil Michailov: Correct. The big difference here is that in big data centers, people also assume big iron. These are racks of equipments and storage devices that are in a lot of cases cost prohibitive. Let’s take a look at a very sophisticated cloud data center like Facebook. Facebook does not buy storage. They’re not a storage customer. They have created that storage layer through software and have enabled a very large-scale and capacity-driven storage infrastructure without buying a traditional storage. It’s all based on commodity infrastructure that is then turned into a storage system through software. There is no argument about how efficient their data centers are, their performance, or their lower cost. I read an article earlier this week that says that Facebook uses open architecture and has saved over a $1 billion in infrastructure cost so far.
There is the notion of rip and replace. If you have a data center that’s been running for a few years, you already have a lot of the legacy infrastructure on premise. You have legacy applications and data. It’s not something that you can easily rip and replace. Quite frankly, you just can’t rip and replace it. It would put you out of business. We focus on pure software enablement.
People are generally reluctant to change. In IT, it’s that much harder because you have this de facto infrastructure and you have all these data and applications already deployed. You can’t just shut them off and move them to something else. We have created a software layer that allows customers to do two things. They can create a unified management framework across that entire storage infrastructure. Rather than managing each one of these EMC or Hitachi devices with their proprietary interfaces, we provide a unified management platform that customers can layer on top of these devices. This allows unification of capabilities in these arrays because inside our platform, we have put caching, snapshot, and quality of service technology. At that point, the storage becomes almost a storage as a service and you drive your workload space on SLAs. You drive your workload space on economics rather than the confinement of the hardware infrastructure you have below it.
The second piece is that customers with that type of data center infrastructures rarely have a single data center. At the very least, they have a DR side. In some cases, they also have a lot of remote and branch offices. Our platform allows them to use the scale-out capability on premise inside a particular data center. They can also span the platform geographically so that instead of having an active/passive data center architecture with an active data center and a disaster recovery data center, they can actually turn on both data centers and scale the workload across both data centers. So, they dramatically improve agility, performance and reduce CapEx, because at this point, both sets of infrastructure are active and productive for them versus one sitting on standby. That scale-out and unified management is what we bring to the enterprise data centers. As they get more comfortable with the pure software approach, they can then refresh their hardware as they go and start to weave in a commodity infrastructure rather than the proprietary monolithic device that they buy today.