HP – stellar cloud or gas giant?
HP has a cloud strategy - in fact, it has several. Does the company need to have clearer plans?
HP had everything required to be a major cloud player. While the majority go for an approach based around "scale out" (the use of a single chip technology where more pieces are added to give more resource), HP is one of the few with different chip and operating system technologies enabling it to work towards a cloud service that can deal with different workload needs.
However, this is looking increasingly fragile, based on various external and internal decisions. For example, HP dumped its own chip technology, PA-RISC, for Intel’s 64-bit chip technology, Itanium. HP’s high-end Integrity servers are all built around this chip – yet Microsoft and Oracle have both dropped support for the chip, believing that Intel Xeon/AMD Opteron chips are where they should put their development dollars.
HP has stated that it will be dropping support for Linux on Itanium as of Itanium’s next generation, leaving it with its own HP-UX operating system as the only viable operating system – but with little in the way of applications and supporting services (such as a database) to make Itanium a viable platform. If HP cannot continue with Itanium, then it has no option but to fall back to a Xeon/Opteron approach, but would then be at a disadvantage to arch-rival IBM, which has the Power chip, the resurgent mainframe and multiple different operating systems that can be used to create a multi-workload cloud environment.
More niche players such as Fujitsu still have multi-chip/multi-operating system plays as well, and even Oracle has a differentiated cloud stack play based on UltraSPARC and Solaris. Falling back to Xeon/Opteron would place HP face-to-face with the likes of Dell and Cisco, though – but is this enough?
HP is one of the later vendors to the cloud party in messaging and capability, but this, in itself, is possibly no bad thing. HP was working on its CloudSystem Matrix, a private cloud approach built on HP’s Xeon/Opteron and Itanium BladeSystem technologies, along with a Matrix Operating Environment and Cloud Service Automation for Matrix.
CloudSystem Matrix takes standard BladeSystem chassis and blades, along with VMware vSphere and LabManager and pulls them all together with a management system built around HP Insight Dynamics to provide a platform for running HP-UX, Linux and Microsoft Windows Server in a consistent manner. Again, though, this is at risk from the uncertainties around the future of Itanium, although the Xeon/Opteron part would still enable a Linux/Windows server play.
HP has also announced HP VirtualSystem for VMware, a series of optimised “blocks” that create a platform for cloud through tightly integrated hardware, operating system and services software, all based on HP’s Converged Infrastructure. The headline is that such a system will reduce complexity.
In the public cloud space, HP has also been busy with announcements. While IBM has been touting its SmartCloud services for some time, HP answered with the HP Hybrid Delivery Cloud (HDC). With cloud data centres under HP’s control, companies can choose to host part of their cloud on equipment owned and managed by HP, using management and provisioning services to integrate these with an internal CloudSystem. It also has HP Enterprise Cloud Services – Compute, a server, storage, network and security platform as a service (PaaS) offering that is paid for on a resource usage model.
Finally, HP has recently announced an offering built around the OpenStack cloud platform, developed by RackSpace. HP Cloud Services has recently opened to public beta, and enables customers to investigate the use of two different services – HP Cloud Compute and HP Cloud Object Storage.
If HP really wants to lead with a message of hiding technical complexity, it should also consider hiding the complexity of its own portfolio. It has too many offerings that seem to overlap too heavily. Its own offerings provide a complexity to the prospective buyer that will only work through a high-touch model, so driving up the cost of engagement and placing barriers in the way of purchase, while its competitors may be able to side step this with simpler messages.
It also has to ensure that its channel base fully understands what fits where and which offering is best for what business problem, and this at a point where HP channel is wondering what HP’s true long-term strategy really is.
And this is not the only problem. HP had been touting the operating system it acquired through Palm, webOS as part of its future server and cloud platforms, yet it has already admitted that webOS will no longer be a part of HP’s future strategic plans. It has announced that it will be bringing out Microsoft Azure cloud appliances – yet another string to an overly strung bow.
HP has confused the market through many of its announcements over the past few years, and CEO Leo Apotheker’s announcements during the last market call have forced many customers and prospects into wondering whether HP should remain as a strategic investment, as a tactical bet or as something to be left well alone.
Just now, HP needs razor sharp messaging – and its cloud message is confusing and overly complex. It needs to heavily compress what it has and ensure that it presents an umbrella where its internal sales, services and account managers, along with its channel can all start and then rapidly drill down to ensure that a customer gets what its business requires.
At the moment, it all appears to be part of HP’s core corporate problem – too much detail with not enough strategic content – an amorphous gas giant, rather than a stellar cloud.