Thursday, September 25, 2008

Is Hulk aka Infiniflex 10000 Delayed or Defunct?


It has been some time (November 2007 – Innovation Day) since EMC let it be known that they were working on a new mass storage device for delivery in the second half of 2008. This announcement was an obvious recognition of the burgeoning opportunities being created by the massive explosion in unstructured and persistent data and the fiscally painful reality that EMC was losing out in some very large storage deals to non-traditional architectures. However, despite expectations Hulk has not yet made its public debut. Not that it has not been shipped, it has, giving the curious and the conspiracy theorist cause for speculation. So let me add to the conspiracy fodder.

Hulk was devised to provide EMC with an answer to those customer challenges created by massive and accretive volumes of images, digital content and unstructured data common in today’s WEB 2.0 environment. The solution demands large volumes of cheap online storage and EMC has realized that their traditional storage portfolio that successfully met traditional storage challenges did not meet the requirements of this rapidly evolving "nextgen" storage environment. In short it was their attempt to answer the Google phenomena.

Hulk is a 300TB, preconfigured, storage building block and despite some innovative mechanical design work, including adding to the height of the cabinet, 300TB raw, while a lot of storage is far from industry leading creating more of a yawn than a WOW factor. However as mentioned it is a preconfigured building block which can be combined to a petabyte scale. As always the devil is in the detail and the question will management transparency and simplicity of use. Ibrix?

From what I understand EMC has overcome the earlier reliability challenges attributed to the new dense packaging, the integration with Ibrix is good to go and API’s are available to facilitate application interoperability. My understanding is that Hulk is but one step away from being a JBOD, however RAID 6 data protection is available. Hulk is available in a hardware only configuration or integrated with Ibrix Fusion software. Maui is not, I suggest, the cause of Hulk’s delay.

So what is causing the delay? My conspiratorial thinking suggests that EMC finds themselves in an uncomfortable dilemma as they rushed to introduce this new, low cost architecture. Their struggle is to define and contain the appeal of Hulk through messaging, positioning or de-featuring to protect their traditional storage portfolio. However in doing so are they creating such compromises that they are failing to create any significant end user appeal and hence sales traction? If this is the case could this signal an early demise of Hulk?

Just a few of my thoughts and discoveries as I ponder on Hulk and what the discussion may be in Hoppington surrounding its fate. I am thoughtful that as Hulk morphed into this marketing dilemma (my assumption) it created confusion, or perhaps fear, specifically, how to prevent the sales team from trying to exploit Hulk for broader opportunities where price is an issue (is there one that isn't?) which would potentially encroach on traditional business.

As always EMC’s provide interesting material for the interested and observant.

No comments: