Converged Infrastructure

[#1]

Guest blog by Brad Parks, Worldwide HP Converged Infrastructure team...

I have a feeling that the word “innovation” will be used a lot this year. I think Kristie Popp captured this well in her blog post where she stated that “If you want to be the best at what you do, you can never sit still”.  A valid point but one worth looking at with a twist and such is the topic of this post.

Those in the world of statistical physics have a name for the random act of not sitting still… random energy in that case is lumped in the category of a disordered system.  While no IT vendor is perfect, I’ve noticed that some have been reduced to wild displays of disordered chest thumping in the absence of real meaningful innovation – often forgetting the most fundamental truth: Our job is to help solve our customer’s most demanding IT challenges so that their IT is best positioned to help drive business success.

At the recent Gartner datacenter conference it was very interesting to observe the discussion around datacenter trends, customer needs, and convergence ... or what they call fabric based computing.  Independent of individual vendor bias, it was clear that customers and analysts are acknowledging that IT sprawl is a reality and simplifying the IT stack by bringing the silo’s together is the only way to free the gridlock.

In fact, even Joe Tucci, the CEO of storage-centric EMC acknowledged many of the same customer pain points as part of a chest thumping display yesterday.  In his remarks to the press he flashed a slide that states ‘IT is at a breaking point’ with 73% of effort going to maintenance and only 27% being applied to new investment.  Joe went on to say that additional customer angst includes too many operating environments, too many vendors, and a lack of control.

I’m writing this not to slam EMC as one might expect but rather to applaud them for getting on board the train.  In fact, I’m impressed they were able to get on board a train already moving at this rate of speed for so long… HP has been talking about IT sprawl and the 70/30 split for over a year (link).  Some of Joe’s slides looked eerily familiar. 

Where the slideware and storyline diverged however was when it came to the payoff.  When the dust cleared from all the marketing fluff it appeared that the big new news from Hopkinton was the unification of two mid-range storage silos into a new multi-protocol storage silo.

Seriously… that’s the proposed answer to IT sprawl and customer angst?  Admittedly I’ve drunk the HP kool-aid, and you may not take my word for it, but having spent 9 years as an IT Director and looking at what independent industry experts are saying about convergence ... I’ve got to think that the future is about unification of infrastructure as a whole rather than unification within a storage silo.

There are different schools of thought when it comes to how best to address the 70/30 ratio, but predominantly HP and others believe we are entering the era of convergence… where the best way to reduce the sprawl and address the customer angst is to break down rigid and expense IT silos to reduce operational costs and accelerate application deployment through a whole new level of simplicity, integration, and automation. 

Even EMC has had to change its marketing approach under the momentum of convergence.  After years of beating on the drum of storage-centricity, they have had to partner with a no-share server start-up to attempt to build a vertical stack and have yet to address the application deployment and automation question. 

Convergence should not come at a sacrifice though … we shouldn’t create new problem ‘B’ as we work to solve IT sprawl problem ‘A’.  Only by having best-in-class technology across individual infrastructure domains and best-in-class vertical integration can we truly help transform IT in the future.  And that’s exactly where HP has been outwardly focused with Converged Infrastructure over the past 15 months (even longer behind the scenes with our Labs and engineering teams innovating convergence).

With over 50% market share, the HP BladeSystem and HP ProLiant has truly proved best in class to tie together networking, compute, and storage. HP Networking has industry leading security innovation with TippingPoint as well as high performance and density from edge to core. 

And then there is HP StorageWorks.  Over the last 18 months HP has delivered a record setting array of organic innovations and has acquired new technologies all with a goal of delivering best-in-class storage designed for best in class vertical integration.  For more thoughts on what’s happening in the world of storage you should follow HPStorageGuy (link) but a few things that spring to mind include:

• HP StoreOnce deduplication from HP Labs delivers leading price-performance and dramatically 
       simplifies backup and recovery with a single software design architected to scale from client to the
       enterprise datacenter.
• HP’s X9000 Network Storage Systems address the ‘big data’ challenge with industry leading scale…
       a 16PB namespace massively simplifies the management of explosive data growth while automated
       storage tiering and a scale-out architecture maximizes productivity regardless of I/O workloads--small
       or large files and I/Os, reads or writes, and random or sequential.

• HP P4000 G2 SANs provide non-disruptive scale-out capacity for virtual environments including the
       only virtual SAN certified by VMware to run as a virtual machine….  Eliminating the need for dedicated
       storage networks in smaller sites.

• And then there’s 3PAR… designed from the ground up to meet the information storage needs of the
       convergence era.  Hyper-efficient, autonomic management, and multi-tenant architecture all contribute
       to the fact that 7 out of the 10 largest global service providers have deployed HP 3PAR Utility Storage for
       their cloud initiatives.

That said, when it comes to truly solving the challenge of IT sprawl, we’ve got to look outside of an individual storage, server, or networking stack.  News flash… HP is not a hardware company and sprawl is not a hardware problem.  The answer requires application oriented management software to pull together virtual pools of resources… enter Insight Dynamics and HP BladeSystem Matrix. 

       Best in class servers, storage, and networking all managed by a single service management
       portal for physical or virtual servers running on any number of different hyper-visor technologies
      all with intelligent power management. 

I’m biased, but that seems like a more ordered and meaningful set of innovations to address the angst that today’s customers are feeling.  Do any weekend statistical physicists out there have a different opinion?

[#2]

Brad,

It is quite amusing to suggest that HP originated the 70/30 IT split discussion “for over a year”. Here is an article, for instance, from Tech Republic from last May discussing how HP copied EMC’s messaging in that regard http://blogs.techrepublic.com.com/hiner/?p=4283. EMC’s (former) subsidiary, VMware, has been messaging about the majority of IT budgets going to “keep the lights on” at least since VMworld Europe in Feburary 2009.  

The theme of your post is that HP is in the vanguard of breaking down IT silos, yet HP continues to propagate them by pitching Virtual Connect to server teams as a way to manage the switches without the inconvenience of network group oversight. EMC, on the other hand, partners with VMware and Cisco to enable a virtual infrastructure platform that both enables and encourages productive collaboration between functional teams.

You conclude by suggesting that the HP BladeSystem Matrix provides an answer to IT sprawl, yet from what I’ve been able to learn, it appears common for Matrix customers to become frustrated with the overwhelming complexity of the solution and instead revert back to using the Matrix as just a server.

[#3]

Hi Brad,

EMC's announcement was heavily focused on its storage core which put themes like convergence (and virtualization and cloud) to the background. EMC is spending hundreds of millions of dollars on the VCE initiative, I wonder how that stacks up against HP's CI group? That "no-share server start-up" is Cisco and we've all seen them enter many markets with incumbents - they don't always win, but they are surely not to be dismissed. HP has an excellent portfolio and one of the strongest mix of solutions built through both internal development and acquisition. That being said, I don't see a virtualization enablement solution from HP that is comparable to either VCE's Vblock or NetApp's FlexPod. It's more than offering the pieces or even bundling, but creating integrated services that make it easier for the customer and channel to tackle that 70/30 issue that we are all well aware of.

Look forward to your response.

Stuart Miniman

Wikibon.org

Twitter: @stu

[#4]

Hey dude… glad I amuse you… reminds me of Joe Pesci however to keep this blog family friendly I will not be quoting any lines from Goodfellas. ;-)

Looking at your reaction it looks like you are suggesting that VCE is more ‘open’ than HP’s approach… “enabling productive collaboration between functional teams”.  I don’t think that breaking down infrastructure silos and productive collaboration between functional teams are mutually exclusive but I’m amused to see VCE positioned as open.  As I mentioned in the conversation with Stu I think most customers at the end of the day are going to run on a combination of physical and virtual with a broad mix of heterogeneous infrastructure and storage technologies and delivery mechanisms.   VCE seems to have taken a page from early automobile production… “you can have any color as long as its black”.

While customers do see value in virtual connect such as reducing cabling by 90% you can also use other networking technologies within the HP BladeSystem including Cisco.  Same is true for solutions like BladeSystem Matrix which do support other storage options such as EMC.  Of course, we would position leading technologies like Virtual Connect and 3PAR storage since they were build from the ground up for virtualized and cloud datacenters but for customers who need connectivity to legacy networking or storage there is an open path to Converged Infrastructure.

Cheers

[#5]

Hey Stu…  thanks for the comments.   Overall I think more vendors focused on convergence is a good thing as it indicates to customers that this is not an isolated trend or play by a single vendor and to your point EMC and Cisco are not to be dismissed in that they respectively have done well in their storage and networking silos.  I also appreciate the comments that Wikibon has made in this space relative to new storage and infrastructure requirements.

While it’s true that they are throwing lots of money at VCE in looking at some of the other items in the news this week it’s not quite clear that VCE knows what it wants to be when it grows up.  The Register said “It's official. The Acadia partnership formed by server wannabe and networking giant Cisco Systems and storage array maker EMC and its virtualization minion, VMware, is now known as the Virtual Computing Environment Company - VCE for short. And it is getting out of the services business and sticking to certification and support of integrated Vblock stacks” and in ComputerWorld EMC execs were quoted “What we're trying to do is position Vblock as an architecture, and VCE as a company that can bring that with one support number”.    Regardless of if it’s a reference configuration or a single SKU, I don’t get that the vBlock is anything revolutionary.  Let’s dissect the discussion a little….   

First, if it’s architecture than it’s tough to see how the roadmaps of 3 separate companies are going to come together to make engineering decisions and tradeoffs that are optimized around the sum of the parts vs. the standalone silos.  Compare that to HP’s Converged Infrastructure approach… our engineering teams are all under the same umbrella with chief technologists from servers, storage, and networking coming together to influence macro level investments to result in a better whole.  Examples of this engineering include products like the P4800 BladeSystem SAN… if we are talking about virtualization enablement this is a solution that physically resides inside the c-Class BladeSystem enclosure but utilizes scale-out storage software to provide industry leading TCO.  I don’t see this type of integration and innovation coming out of a loose partnership like VCE.

Second, if it’s a certification and support engine then I’m not sure of the customer experience.  VCE has stated that its sales arm is still made up of the siloed EMC and Cisco sales teams but I guess they come together into a single support entity?  When you peel the vBlock onion apart you still have independent Cisco gear and EMC gear so I’ve got to think level 2 and level 3 support is still going to be diving back into a silo and there might be some finger pointing within the VCE family.  Troubleshooting and support is always an interesting challenge but I’ve got to think that HP has a leg up with an award winning services arm, 30 global multivendor labs to assist in complex support issues, and over 175,000 channel partners worldwide all of whom have been providing a single point of support for decades.  Extending beyond support to other services we should look at HP Enterprise Services and their robust capabilities to help customers with application transformation efforts, information optimization, and datacenter transformation.

Lastly, when looking at virtualization enablement I think it’s critical to look at what customers are actually deploying in their datacenter… what is it that actually answers the business problems?  I’d posit that it is the applications which address the immediate business need… businesses are looking for a better car, not just a better engine.  In most datacenters those applications are going to run on a combination of physical and virtual servers, across a mix of heterogeneous operating systems, across multiple hypervisor technologies, across storage devices that comprise network and direct attached storage, and are going to be consumed through a combination of in-house dedicated IT, in-house clouds, outsourced, and hybrid models.  

Thinking about the application angle for a minute I think that we are looking at application enablement rather than just virtualization enablement… take the Cloud Maps and application templates within Insight Dynamics or the set of Converged Application Solutions that HP announced this week.

HP is in a very good position to deliver unified and application-oriented solutions, with a single point of service & support, built on an integrated architecture which can span ALL of these needs.

Cheers!

[#6]

What I don't really understand here is you start off talking about convergence, but then when you get to HP StorageWorks, the only converged thing I can see is the name.  You mention StoreOnce, the X9000 (IBRIX), the P4000 (LeftHand), and 3Par.  You don't even mention EVA, or the P9000 XP arrays, or PolyServe, or the X500/X3000 NAS that runs Windows Storage Server.

If anything, you've got more storage technologies than EMC, or if not, an equal amount of them.  What on earth are you going to do with all of those?

[#7]

Seems like HP is paying a lot of attention to the competition.  VCE is very confident in where we are heading, and the demand in the marketplace validates that.  Your assumptions are amusing to read but pretty far removed from reality, and with as many HP customers as we are talking to they aren't buying your narriative either.

I'd like to tell you not to worry about VCE, but the truth is that you need to worry about us.  HP is the low-hanging fruit right now, full of customers who are actively looking for a better way to run their infrastrucutre.  For all of your talk about one company, one support, everyone knows better.  HP server sales and HP storage sales and HP network sales are completely "siloed" entities, in some ways moreso than VCE.  When I was an HP customer we had separate VARs for each product set, which ment separate POs and separate sales relationships.  What part of that is converged again?  Same goes for support.  We still had to open a ticket on storage, a ticket on compute, a ticket on network and a ticket with VMware.  Where is the value there for the customer who is counting on the unification of their infrastructure to make life better?  Just because the same name is on all of the kit doesn't make it unified or converged and I'd wager that most HP customers understand that pain.

I work for VCE, specifically in the SP and Vertical Solutions group, and I am a reformed HP bigot.  I'm here now because I believe there's a better way to help Service Providers and enterprises deliver services, and I believe that VCE can deliver that.

[#8]

Hey Roger… you’re right that we have a broader and more complete storage portfolio than EMC.  Great point :-)  

In all seriousness we do have a lot of solutions but customers at the end of the day have a lot of different requirements.  I think convergence is a journey and while some are ready to deploy fully automated and application oriented service catalogs, others are still trying to simply standardize on common building blocks to improve efficiencies.  Regardless of where they are on that journey we’ve got an opportunity to help them reach their goals in a way that is open and extensible.

The other solutions you mention are also good examples of convergence… EVA and XP are core elements of the portfolio for enterprise applications and both are manageable as part of the HP Storage Provisioning Manager within BladeSystem Matrix.  Additionally the X-Series  of NAS solutions from low-end to high are all running on BladeSystem and ProLiant hardware meaning they can be managed through the Insight Control portfolio with the same proactive service notification and remote management capabilities that HP server customers rely on every day.

[#9]

Hey Jeramiah… of course we pay attention to the competition, it would be silly not to.  I’ve got to point out that you obviously pay attention to the competition as well… evidenced by the fact that you are finding a voice here on an HP blog.

During the client server era enterprise IT was pretty fragmented with server, networking, and compute largely managed in isolation.  This was also reflected in the technology industry with vendors like EMC and Cisco focused exclusively on storage and networking respectively. HP’s internal dynamic naturally also reflected this split.  You’re right to point out that that having a separate relationship with a networking vendor, a storage vendor, and a server vendor wasn’t always the best thing for the customer… regardless of the badge on the kit.

Over time though that has created a new problem… IT sprawl.  With advances in automation and widespread adoption of virtualization we are moving into the convergence era.    From engineering to sales and service all of the wheels are pointed in the same direction and that is the Converged Infrastructure.   Even the leadership ranks at HP has had an influx from companies like EMC and Cisco… folks who see that HP is in the best position to truly transform the datacenter of the future and help customers improve efficiency and agility.

I’m also really excited that the convergence is not just at the hardware level.  With the intellectual property that HP has from datacenter design to software and application automation I think that this is an exciting time to be in this space.