Re-Shaping of the Data Center: The Scoop on the 3Par Bidding War

This morning HP announced they were trumping Dell’s bid for 3PAR, offering a whopping 33% on top of Dell’s 85%+ premium. I’ve read lots of chatter about why, but I think much of the analysis misses the mark.  I thought it would be worthwhile putting the deal in historical context.   

First, let’s talk a little bit about data center and storage history.  Storage many years ago used to be housed with the CPUs.  The large vendors shipped computers with high end processors, packed with disk drives and called them servers.  As data storage grew faster than compute, the industry began to decouple storage from the compute, giving birth to companies like EMC. 

EMC and Hitachi Data Systems operated in the high end enterprise segment with very large storage arrays designed for high performance, availability, and reliability.  Because of how critical and difficult storage is, very few of the server makers chose to wade into the market.  In fact, most compute players would regularly OEM product from the specialized storage makers.  This led to a very happy symbiotic market with clean lines where everyone knew their place, and each of these vendors in fact OEM’d one another’s products.  For example, Sun and HP each resold HDS’s products.  Dell resold EMC’s products.  Brocade was built almost entirely on a channel model.  

A couple of moves really changed this panacea.  First, EMC got a hold of VMware and virtualization subsequently became the hottest trend in the data center.  This pulled EMC into the server side of the market and led them to rapidly expand beyond storage into systems management, software, and other layers of IT spend.  Second, Cisco announced they would be entering the high end server market.  This clarified their growing ambitions from dominating the router market into the compute part of the IT spend. Cisco announced a JV with VMware and EMC to complete their product vision late last year. Third, Dell bought Equallogic and HP bought LeftHand Networks, both signaling a movement towards owning IP for storage (albeit the mid market).  Dell had been partnered with EMC going back to 2001 and was a meaningful channel for EMC’s mid range products.  Very quickly everyone got a wake up call that their place in the stack was not secure.

So what’s happening now?  Every major data center platform vendor sees two major trends going on.  First is the rise of the dynamic, agile data center within enterprises.  This requires being able to spin up resources – compute, network, and storage – automatically in response to business demands.  Second is the eventual move of the data center to private and public cloud offerings.  In this model, the vendor no longer sells equipment to the enterprise, but assembles and runs all the parts as either a dedicated or shared service.    

In order to fulfill this vision, you need all parts of the stack working together seamlessly.  This is where 3Par comes in.  3Par was born during the great storage gold rush of the early 2000s.  Bringing their product to market took over $200MM in venture capital, including some recaps along the way.  They were one of many startups that were funded to build flexible, modular, high end systems, but one of the few to survive.  An enterprise’s lifeblood is storage and they would not trust startups lightly.  This required high burn to build the technology, and then high burn on the sales side to succeed in market.  To 3Par’s credit, they managed to get public and raise sufficient capital to sustain themselves to critical mass and profitability.  And now they benefit from scarcity value.

Looking at the landscape, 3Par is the only real alternative to EMC and Hitachi in terms of high end storage.  EMC has its own ambitions for data center dominance, while HDS is part of a much larger conglomerate.  If you believe you need to own storage and server, both to fulfill the vision above and to avoid partnering with a competitor, than 3Par is the only place to get this type of deep high end storage technology.  Given HP and Dell have a much larger sales channel than 3Par, these guys can immediately double, triple or quadruple sales from 3Par products overnight once it is part of their catalogue.  Both reasons afford the premium we are seeing.

Going forward I’d expect to see more data center consolidation.  There are some major battles brewing as companies compete to own the enterprise! Network Appliance has long been rumored as a fit for Cisco.  Plenty of other combinations make sense as well.  It’s clear to me, though, that the march is towards creating end to end solutions and masking complexity.  Should be a fun next few years to watch!

Tagged , , , ,

11 thoughts on “Re-Shaping of the Data Center: The Scoop on the 3Par Bidding War

  1. Alan Quayle says:

    Hi Amish, great article. You’re right the move is to treating computing, storage and connectivity as simply infrastructure bought on demand to meet an applications needs. With HP (3Com), IBM (Juniper) and Oracle (Sun) delivering impressive verticals with DC and on-demand dominance as 2010/2011 objectives, it will be interesting to see if Cisco maintains its DC edge, and also if the web-service providers (Google and Amazon) break out of their developer niche into mainstream enterprise DC services. Its definitely something that’s going to play out over the next 5 years given the security and availability concerns remain significant hurdles hence continue to slow mainstream adoption.

    • Amish says:

      Hey Alan. It’ll be an interesting battle between the large enterprise guys who understand the security issues and the web services guys who focus on innovation and cost.

  2. Mattthew Barr says:

    Hmm.. I’d add Compellant as an option, and there’s also IBM’s XIV. But the options aren’t huge. HP does have the EVA & MSA lines, but I don’t think they’re nearly as flexible.

    • Amish says:

      They can’t own XIV (and not sure it scratches the itch). Compellent is a pretty reasonable potential, but it’s designed as a mid-range system with incredible features. 3Par is designed throughout to be rock solid and high end, so I don’t view them as interchangeable. Given HP already has a solid mid-range, as does Dell on IP side, the real hole is at the high end (which subsumes the utility model). IMHO.

  3. jp says:

    Excellent summary of the Hp/Dell/3Par situation. I’ve read in several places that Dell “needs 3par” more than HP. Untrue, HP needs 3 par as much if not more. HP’s EVA is not up to the challenge from the higher end EMC. I don’t think Mr Donatelli wants to resell another co’s gear. He was an EMC guy, he wants to go after them now that the non compete is done.

    • Amish says:

      Functionality and product-wise, I totally agree. But HP already has a strong “relationship” at the enterprise level. Dell is a “supplier”, and is trying to build a deeper relationship. In that context, there is nothing more sensitive than storage and getting 3Par for Dell brings a new level to the relationship they absolutely want.

  4. Ravi says:

    Amish,

    Very smart and clean analysis on this story. Thanks for the same!

    Regards,

    Ravi

    • Amish says:

      Thanks Ravi. I have had the pleasure and pain of watching this segment for a while…. 🙂

  5. […] Jani is a Managing Director at FirstMark Capital. This post originally appeared on his blog, and is reprinted with […]

  6. Avi Deitcher says:

    Amish, nice review, be great if you could get someone from Dell or HP to comment on it (sure!). You are right that big companies won’t trust their storage (or at least the data) to startups, especially information-critical firms like financials and health-care (although seeing what they do with the existing companies like Siemens and IBM makes me wonder why not). I spent a decade on Wall Street IT

    However, I think many enterprises will avoid 3rd-party provided clouds entirely, whether private or public. First, many of them have the economies of scale to do it on their own. Second, many are loathe to trust their data to anyone else, even a large firm. Finally, there are regulatory reasons for needing control of the data, e.g. PCI-DSS or HIPAA, which require actual physical separation and control. Those can be achieved via private cloud providers, but they are much more expensive, and quickly lose the entire concept of cloud value.

  7. Avi Deitcher says:

    Sorry, previous comment got cut off in the first paragraph. I meant to say:

    I spent a decade on Wall Street IT, and can understand why these guys would not want to send business-critical systems outside. While I am comfortable with it now, I would not be comfortable doing it if I were still there.

Leave a comment