HP: The Legos of enterprise IT

Security breaches from lost laptops, consumerization of enterprise IT, and a slowed economy have all been factors forcing current CIOs and IT staffs to find ways to cut costs and increase efficiencies. Virtual desktops while often one of my primary topics has become just one of the many pieces of the puzzle. The others include but are certainly not limited to the server infrastructure, networking equipment, enterprise storage and all the components that allow them to run, both physically with cabling, cooling and power, but also virtually, with management consoles and user interfaces.

Over the past two years, almost every major vendor in the IT industry has come up with a new fancy name to say that they are combining all the components needed to run your enterprise. Acadia was a collaboration from EMC, VMware and Cisco which later became VCE, Netapp was soon to follow with the Flexpod. Both of these are great solutions and are clearly geared at the large enterprise that buys in bulk. They both allow you to control the storage as part of a solution that includes the servers and even includes hooks into the virtual infrastructure. However, they also come with a cost, and a large one since you have to buy the units as a full configuration all at once.

Next Big Thing

This is where HP may have found a way to crack the preverbal nut. Over the past few years HP has made multiple strategic acquisitions that have left many of us in the field wondering why they have done what they did. Lefthand instead of Equallogic. 3Par instead of Compellent. Then there was the release of StoreOnce from HP labs, this one made sense since deduplication continues to be a hot topic. Then there was Ibrix, a clustered scale out file system. Was HP just looking to buy up the next big thing but none of them ever really took flight or did they have a grand plan. Well it appears the grand plan may finally be emerging. The first of three sessions presented to the contestants on the Blogger Reality Show was on what HP has coined “Converged Storage” It definitely does not have the ring to it like Acadia or Flexpod, but it does tell a story. Converged Storage is the culmination of those purchases, a scale-out and scale-up architecture that can not only be purchased as a total solution and provide end to end technology but also an architecture that can be built over time. More often than not, the reference architectures written by the likes of VCE, NetApp, or EMC and often HP revolve around the largest Fortune 500 companies. Reference Architectures for desktops are built for 10,000 seats, and the server platforms are for hundreds and thousands of servers. Where does that leave the mid-size company with 250-2000 users and 20-150 servers?

Building Blocks

The best part of the Converged Storage platform in my opinion is that it can fit both the large enterprise as well as the small to medium business. HP’s reference architecture for VMware View is in fact built to house the “micro-branch” through the enterprise. Each of the segments is built like a city built from Legos. First you build the lumber yard and hardware store to keep all the pieces you need. An initial storage component like the p4000 VSA will give a micro-branch a storage place that works just as well as a foundation. Next you would move on to build some utilities; your power is driven by some industry leading x86 Proliant Servers. As you city grows you needs to have more multi-family (tenant) living and you move into a blade chassis. The c7000 has for years been the go to blade chassis for the mid-sized market. We are starting to see come competition from Cisco, Dell, and IBM but most of the blade chassis seem to be the c7000. Even as we have built the city up it still seems very segmented. This is where HP gets interesting, for the customers out there that already have an equallogic array or a c7000 chassis you can simply add to them and start integrating all your components. HP released the p4000 storage blades that will allow you to connect to Storageworks 600 disk shelves. These blades take up space in the c7000 chassis and move the hardware store from the other side of town to the ground floor retail in your apartment building. While comparing an enterprise solution to a kid’s building blocks may seem like a jump, we are still just trying to get all the pieces to fit together. In comes the planned community, the HP p4800 gives you all the parts to your infrastructure in a pretty package, with blades, storage, and connectivity.

Citizen Patrol

With the whole city built we can now look at how it actually helps the IT staff that is trying to fix today’s enterprise issues. With the ability to scale out, an IT department can now buy a piece of infrastructure as their old equipment runs out of warranty or lease and start to see the building blocks come together without wasting the precious money they are allotted. One of the first use cases for this may very well be the virtual desktop space. The reference architecture for Client Virtualization with VMware View gives some great examples of how we can build VDI deployments for smaller, sub-1000 user, deployments with the p4800 configurations. Storage speed, bandwidth requirements, and server performance are all taken into account in these configurations. The used of the blade chassis backplane allows for near instant access to the storage. By adding in an additional chassis and more p4000 storage blades you can create a fault tolerant desktop solution with a single management console. These same solutions could be created with the FlexPod and VCE solutions but starting so small is a clear differentiator.

The crazy neighbor

HP has managed to integrate Proliant servers, Lefthand storage, and their core Flex-10 switching, but it seems as though 3Par was left out of the equation. HP will integrate the 3par solution for the high I/O needs into an architecture, however HP is still lacking the same component all the major vendors are, which it true cross platform integration. If a c7000 chassis was placed in a headquarters with a 3Par array as the storage, and the p4800 with Lefthand at the branch offices, and replication could occur between them along with centralized management, HP would be a clear winner in the “Converged” space moving forward.

6 thoughts on “HP: The Legos of enterprise IT

Add yours

  1. Hello Michael

    Interesting article, just some questions

    1)what solution you would pick and why?
    2)if none is the answer for (1) what could be the best solution approach to follow and why?

    Could you include a table of comparision among these options?

    Best Regards!

    1. One of the great things about having so many solutions out there right now is that we have lots of options. It really depends on the size of environment. I think it is rare that an existing company needs to do a full rip and replace. I think the ability to use the existing and just grow with the HP side is a great option, if you already have HP in place. There are clearly some advantages with some of the other storage vendors in respect to tiering that is not as baked with the LeftHand but if it is not needed the add that to existing servers and start moving forward. If you already have the c7000 it is also a great option on the storage add-on.

      As for a chart on the comparisons, I think the reference architecture does a great job of showing when to go to what level depending on office size. I am not sure I have the cycle right now to come up with a great comparison chart that is any more valid. http://h20195.www2.hp.com/V2/GetPDF.aspx/4AA3-4920ENW.pdf

  2. Mike,
    I like how you structured the post, getting my attention with the security stuff..nice..actually raises my questions on the whole security aspect of all this more now…

    Like what security measures does hp utilize in a virtual environment other then VMwares offerings? Think i just found something to research and write about thanks!

  3. Mike,

    I agree… replication between their classes of arrays is the holy grail for HP and its certainly something I think that customers will want. It will make migrations and disaster recovery a snap.

    It is also good to hear that they are working towards a fully integrated management platform to manage all of the array lines from a single pane of glass.

  4. Mike – I was in So Cal last week, not far from Legoland – dang, I really wanted to go after reading your post but heck, I was staying right at Disneyland and didn’t have time for that either. Thanks for making time to be part of the contest. Looking forward to meeting you at VMworld!

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Proudly powered by WordPress | Theme: Baskerville 2 by Anders Noren.

Up ↑