November 28, 2012
I’m heading to Europe for HP’s Discover event and the conference has me thinking about the last Discover event I attended in Las Vegas earlier his year and HP’s awful few weeks around the Autonomy debacle. Alongside my theme du jour of traditional enterprises (and traditional vendors) being disrupted by new, more flexible and adaptable firms, has me thinking about the long-term implications of industry upheavals on HP, and what we can expect to see announced at discover. Over on GigaOm Barb has an interview she did with HP’s GM of Cloud Services recently, Biri Singh. In the interview, Singh is bullish about HP’s chances to shine in the new and growing market of cloud services. Let’s take a look at that.
HP is yet to deliver a fully-featured offering covering both cloud compute and storage. Indications are that the work that HP has been doing on OpenStack will see them announce the compute piece of their offering at Discover, this will go with the storage and CDN components they already have. It is important to remember that this week also sees Amazon Web Services hold their event, and it is important to contrast the maturity of these two cloud plays – AWS is the undisputed leader in cloud infrastructure – their offerings are both broad and deep and the basic storage and compute building blocks where available from them many years ago. Given this fact, it’s hard to imagine how HP can win market share with a somewhat rudimentary offering. But to understand why HP shouldn’t be written off, at least in the short to mid-term, it’s important to think about Singh’s contention that HP’s staple customers (the largest global businesses) want guaranteed reliability backed by contractual terms:
[customers want] raw service level agreements (SLAs) in which we’ll guarantee you this much performance or will pay you for it. We think our SLAs will emerge as a differentiator for any cloud vendor
Now I’ve gone on record as saying that SLAs are an incredibly blunt instrument – any organization that relies on them as their guarantee of service is destined for tears. After all most SLAs only pay out the aggregate value of spend over the outage period – small comfort to multi billion dollar enterprises that rely on technology to deliver their gazillions. Add to that the fact that, as Netflix’s Adrian Cockcroft famously espouses, in a cloud world users should be architecting for failure, and you have a couple of reasons that would, at face value, indicate little room for HP to move.
But…. enterprises doesn’t work like that. For all the cloud evangelist rhetoric about agility, flexibility and light weight services, large enterprise is big. It is slow. And it demands rock solid guarantees. It quite simply cannot comprehend a “design for failure” mentality and a core assumption that discrete parts of the operation will suffer outages over time. Rodney Rodgers, CEO of Virtustream said it best in a post recently when opining on the fact that enterprises are not, by any stretch of the imagination, sexy:
I am a large enterprise. A vast array of technologies – with varying degrees of purpose and tenure – are required to run me. I have a complex, heterogeneous technology landscape. I have hundreds of millions of dollars of technology assets deployed on laddered refresh cycles through a variety of delivery models, all supporting highly differing divisional requirements across a number of development platforms. I employ thousands of developers, architects, and technical managers. I own just about every form of software known to man.
I am a large enterprise. I am not sexy. I am complicated. This will not change.
Large enterprises tends to have a risk averse and coarse logic to their decision-making process. When it comes down to it, a technologist within a large enterprise is unlikely to recommend her organization rely on a vendor when another vendor exists with a better promise of reliability. enterprises live on individual self-protection, what enterprise IT staffer is likely to run the risk of looking bad in the event that their lower-availability cloud service goes down? This is why percentage points (generally shown in 10 point font on 30 slide PowerPoint decks) are critical. Singh mentions this fact in his GigaOm interview when he says that:
[HP’s] SLAs for object storage and CDN offer 99.95 percent availability compared to 99.90 percent for Amazon storage and CDN. HP recognizes one failed instance as “unavailable” while AWS says all running instances have to be without external connectivity to be categorized as such
Of course this isn’t simply a dissertation on HP cloud versus Amazon cloud. HP is famously “all in” with OpenStack and hence competes with it’s OpenStack community members for market share. While the OpenStack community does an awesome job of keeping a relatively united front when talking about the big aspirational topics, when it comes down to sales, it’s a safe bet that they watch their own backs, and the moves of other players closely. We’re starting to see a more open articulation of differences between the various OpenStack players. Singh isn’t shy to hold out HP as the “trusted OpenStack provider” saying:
…it’s the “OpenStack Plus” argument. HP thinks its enterprise software know-how… will make its OpenStack shine above the others. And, Singh said it’s done a ton with OpenStack since it started working on it last year. The SLA argument holds true here too, he said. According to HP, Rackspace offers 99.90 percent SLA for storage and CDN compared to 99.95 percent for HP. Finally, HP being a hardware company still, hopes to press that advantage as well by “fusing” OpenStack into its converged hardware.
Again a reflection on the reality that HP sees with its bread and butter customers who, for better or worse, would rather obtain the bulk of their technology from one vendor. Customers who spend equal time perusing the contractual terms of any arrangement with the technical specifications of the solution. Customers whose careers depend on making the right choices for their employers.
When taken in this light, and in stark contrast with those who proclaim the end of companies like HP is nigh, it’s not hard to see their proposition being compelling to some companies. Yes AWS and other vendors are more innovative. Yes, one can argue that a technology strategy that focuses on SLAs as a guarantor of reliability is missing the point of the cloud. Yes, one doesn’t “need” the converged software/hardware and “iron clad” guarantees that an HP solution might deliver. But none of this matters. Enterprises make buying decisions for a multitude of reasons, and flexibility isn’t high on that list. Acquiring the very latest in functionality doesn’t rate very highly either. In this environment, an HP badge is worth several generations of technology innovation, a fact that HP and their ilk should be very thankful for.