On June 7 at HP Discover in Las Vegas we announced a new platform called HP VirtualSystem. I had the good fortune to be able to do the keynote presentation introducing VirtualSystem, and last week did that same keynote presentation in Hong Kong for our partners and customers in Asia. This week I’ll be presenting again at HP Storage Decisions in Chicago and elsewhere. I’ve signed myself up as the unofficial “promoter-in-chief” for VirtualSystem because I think it is going to be the most impactful thing we introduce this year, and it is one of the best examples of converged infrastructure that has emerged to date. Let me share a couple of key points.
I believe that in 2011 we are just entering what I call "phase 3" of server virtualization. "Phase 1" was when organizations started using VMware and Xen for test and dev applications, but not for mainstream use cases where they might get into trouble if things went wrong. In "phase 2" we saw ubiquitous use of server virtualization for business-critical applications. “Phase 3” is where virtualization moves into the infrastructure, and we build all kinds of other capabilities on top of it. Virtualization becomes the launching pad for cloud, big data and IT as a Service in “phase 3”. In reality, we are seeing virtualization become fundamental to IT infrastructure not just in server technology, but in all parts of the infrastructure. All new storage, networking and application technologies are now built on virtualization technology and concepts. In HP’s storage business, this is exactly what we have done - all of our new storage technologies are built this way. In fact, since the emergence of EVA (Enterprise Virtual Array) ten years ago, HP has led this movement "toward virtualization" inside, and this has been a key reason for our success. But "phase 3" is going to be about an entirely new level of reliance on virtualization and is a core technology within the infrastructure. In "phase 3", all the parts need to work together more seamlessly than ever before.
To get to “phase 3” we need to fix some of the problems that customers have experienced in "phase 2". It turns out that many of the challenges that customers have run into in server virtualization deployments have actually been storage problems. Think about it this way: if you take a physical server that is using 20% of its capacity, and you virtualize it and consolidate such that it is now using 80% of its capacity, you run the risk of overloading the back-end storage systems. If you don't change the backend storage infrastructure that is supporting that virtualized server, something is going to break. Customers that turn on mobility or "motion" technologies on their virtualized servers often find that the backend storage is not able to keep up. The LUNs and volumes are mapped in a fixed manner and the storage architectures they have in place are often not agile enough to deal with these new dynamic workloads. What often happens is customers end up over-provisioning their storage so they can avoid hotspots and performance problems, and all the money they saved by consolidating their servers gets spent on deployment of additional storage capacity.
One way to address this is to use storage technology that was designed in the virtualization era, and is agile enough to handle anything VMware, Microsoft or Citrix virtualization technologies can throw at them. Clearly, this is what 3PAR and LeftHand were designed to do. Another element of the solution is to use a virtualization "system" that is tightly integrated from end to end, across the server storage and network layers. This is what our new HP VirtualSystem is designed to be.
There are three VirtualSystem models - small, medium and large - and they will be orderable starting in Q3. VS1 is based on ProLiant servers and P4500 LeftHand storage. P4500 software running on a ProLiant, so the infrastructure is consistent from top to bottom. VS2 includes BladeSystem servers and P4800 (LeftHand running on a Bladesystem). VS3 is BladeSystem and 3PAR. These systems are rated for 750, 2500, and 6000 virtual servers respectively. All can support both VMware and Microsoft server virtualization, and VS2 can also be used for Citrix client virtualization. All of these systems take advantage of HP VirtualConnect networking, Insight Control management, and TippingPoint security for threat detection. They can each be ordered with a single part number. A complete system for virtualization.
If this sounds to you like an EMC vBlock or a NetApp FlexPod, that comes as no surprise, because those are probably the closest competitors to HP VirtualSystem. VirtualSystem has a variety of advantages, but the most important ones are:
The last benefit is a big one. Because we construct our "systems" out of hardware and software building blocks, the hardware components of VirtualSystem are actually identical to the hardware in HP's CloudSystem. This means that in the future if you want to convert your VirtualSystem to a cloud, is a software upgrade. Bring in the software components of HP CloudSystem (CSA, CloudSystem Matrix) and you're there. This is entirely unique in the industry today.
We encourage you to share your comments on this post. Comments are moderated and will be reviewed and posted as promptly as possible during regular business hours.
To ensure your comment is published, please follow our community guidelines.
