Ed note: The following are my working notes, transcribed from the recording of an interview by Wikibon co-founder David Vellante with two top HP storage executives at VMworld 2010 on SiliconAngle.TV. This is not intended as a world-for-word transcription but is close to that and is posted as source material for others interested in this specific area. Note that speakers are identified by their initials in bold.
Participants:
- David Vellante, co-founder and CEO, Wikibon.org
- John Furrier, founder of SiliconAngle.com
- David Robertson, Executive, HP StorageWorks
- Paul Perez, Executive, HP StorageWorks
DV: I want to talk today about the transformation of HP's storage business. These kinds of big transformations generally take awhile, and you guys have been at it for a number of years. Dave, What's new at HP StorageWorks?
DR: We are focused on 3 areas: The most exciting is the StoreOnce technology that we've been working on with HP Labs. An end-to-end data dedup solution. The first instance we brought out is a disk-to-disk backup device. We are excited about the price/performance and end-to-end capability of that software. Then we are also focused on our P4000 G2 SAN (formerly LeftHand) product line, our iSCSI scaleout architecture. We introduced the 4800 in June, a virtual machine for the virtual desktop preconfigured with a blade architecture. Then the last thing is the StorageWorks X9000 Network Storage System (NAS) product line of single namespace scaleout file-based storage. Very excited about the acceptance of that in the marketplace.
DV: HP Labs is a tremendous resource, not always monetized effectively. HP completed the acquisition of Storewise, so a key area, you came up with StoreOnce. It is not a product limited to a single use case. It can be applied to different areas. Paul, can you talk about that?
PP: Essentially what we see is an end-to-end secondary storage division where customers can optimize that backup stream whether you are starting from a virtual machine, a backup to disk appliance or a media server. We can capture that stream, dehydrate it, & from that point on through a LAN into a central store.
DV: The problem with a lot of the products we see is they are stovepiped. There's one technology here & another there & another in a third area. So when you start moving things around the network you have to rehydrated them, and that causes clogs in the network. So HP's vision is really to put the IT anywhere.
DR: Dehydrate it once & because it is the same algorithm, same process, same software you can move it around. It's all consistent in all these use cases.
DV: What do you see as the mix going forward?
DR: We prefer to innovate ourselves wherever we can, but if we have gaps & there is good technology in the market, we aren't opposed to acquiring, also.
DV: Is it a misapprehension on my part that HP Labs has not been as productive as it might have been?
DR: I think the relationship between HP Labs & Storage Works is one of the strongest in HP. StoreOnce is a great example of a real technology that is leading in the industry developed jointly by HP Labs & us. There are other things I am not ready to talk about that have that kind of impact in the marketplace going forward.
DV: How about the convergence story. HP has talked for years about leveraging its server piece & you have done a great deal with the server portfolio. It seems different now. You've always had a really strong server business. Bringing in Donatelli (sp?), who has real credibility in the storage business, makes you stronger. Do you think you can leverage that to a greater extent than you have in the past & why?
DR: The simple answer is yes, we can. If you look at the shared components or shared infrastructure, many of our products r built on industry std or our blade servers. We aren't having to engineer t hardware component we are running on now, so we can focus on the software, differentiation, leveraging HP supply chain across the various businesses including the PC business to really take advantage of the engineering, purchasing power & relationships we have with strategic suppliers to be on t leading edge of that.
DV: You'r supply chain is enormous -- $50B+. How do you leverage that as a tactical & strategic weapon in the marketplace?
PP: If you think of converged infrastructure as a development community, we derive benefits from being a member of that, but we also give back with a slingshot effect. For instance we did a conversation at VMworld around the storage provisioning manager. Because we didn't have to work on the hardware & platforms, we were free to innovate on the software level. We're doing it on the storage domain & using it to bridge across server, storage, & networking to accelerate automation.
DV: Talk about the execution of the vision of a common platform, modular components & then changing the personality block or file through software. Where are you at in the maturity model?
DR: The platform is there already. Many products run on our standard servers. A good example is the Cloud Start product we announced here, where storage, networking, services, software & hardware all brought together to make it much simpler for customers to start a cloud from all aspects. You don't have to bring in 3 different partners to get an end-to-end solution. So we see storage as part of our end-to-end solution rather than here's a server, there's a networking connection, and over there is a storage device, go figure out how to bring them together. If we can integrate all these components into 1 answer, it is much easier for the customer. We bring the integration to the customer. Historically the customer, a reseller or some other partner has had to do that. Not to say they can't, & certainly there is a lot of value add from our partner community. But we can make it easier for our partners to bring these solutions to their customers as well.
JF: The mainstream press is analyzing why storage is so hot & coming to the conclusion that it is the cloud. Specifically the analysis of storage as it is in this environment. Now it is the lynchpin of cloud virtualization. What does that mean for HP, which has a breadth of resources? Where are people missing in the analysis?
PP: Dave asked about vision as well. In the plate tectonics, the 2 big shifts in the market are the unrelenting growth especially in unstructured data & acceleration in server virtualization. The fault line where they meet is storage. So the biggest barrier to adoption of private cloud in virtualized environments is the cost & complexity of shared storage. At a high level we are making that simple, cheap & turning an exposure into an acceleration devices.
JF: And data playing a role in mobile devices.
PP: Absolutely. What happens is the old model was taking data and piping it through a network to be available to applications. The new model is bring the virtualized applications to co-reside with the data & run it there. It's a lot more efficient & a lot higher performance.
JF: It is a complete rethinking of the network. The enabler of that is virtualization, & I think you are saying storage is fundamentally like a server & networking in one.
DR: That is 1 way to think about it. Bringing these elements together with a simpler management paradigm, which we've done, makes it much more cost effective. Another thing to think about is historically we've had storage administrators, network administrators, server administrators, database administrators, & in the new world it becomes an infrastructure administrator at some point. Where its a different view. Focusing on the policies, use cases, how to protect the data, how many copies. Some customers are realizing this and saying our IT organization needs different skills.
JF: Security has been another adoption issue with cloud.
DR: Security is important & HP touches security in many forms, we have a very broad security capability set . The tipping point software we got with t 3Com acquisition, for example, is one of t leading network intrusion softwares, so that is a great example. We take security seriously. It is important, and every customer cares about it. In t financial services world they care a lot about it, but everyone cares abt it.
JF: What's t general philosophy with open source?
DR: One fundamental building-block is an open standard architecture so customers aren't locked in to one particular answer or vendor. The legacy has to work with the new. You can't tell the customer that they have to throw away all the stuff they bought over the last 5 years to use the new stuff. For one thing they can't afford it. So let's virtualize it to transition to the new world, & over time we move to the new. That's our view of it. But it has to be based on open standards. You can do it proprietary but usually the customers won't accept it.
That is a question we get a lot at HP. Customers say, “We can buy everything from you, but does that put us at risk?” I say absolutely not because we're based on open standards, so you can replace whatever you want along the way. You can't put a Dell blade in an HP blade cabinet, but you can connect other storage to our servers & vice versa.
DF: You have Procurve, 3Com, how is that converged organization going?
DV: Certainly the engineers more excited than they have been in a long time. We investing more in R&D. Engineers want that. And we have introduced strong accountability. Dave Donatelli has brought us strong direction with more money & here's what I will hold you accountable for delivering to the market with that money. As an engineering team, that's what you want.