HP has announced Helion, its next-generation public cloud offering. Based on the Icehouse version of OpenStack, Helion includes substantial additions from HP, including a PaaS layer based on Pivotal CF (Pivotal Cloud Foundry). HP also announced a free community edition that is limited to thirty servers. This will prove an interesting test of whether any vendor can build a credible competitor to Amazon Web Services around OpenStack.
HP Helion Overview
As mentioned above, Helion is based on OpenStack. It comes in two editions, which are summarized in the diagram below. The short version of the story is that HP is making a community edition of Helion that offers some of HP’s value-added features for free for environments that have fewer than thirty servers. Larger environments and environments that require enterprise-grade capabilities need the full HP Helion OpenStack edition.
HP Helion Value Add to OpenStack
Helion is not just HP’s distribution of the Icehouse release of OpenStack. HP has made considerable enhancements in the areas of lifecycle management, integration, extensions, certifications, and indemnification. The relationship between the core of OpenStack and the entire Helion offering is shown below.
Helion is not just an IaaS cloud offering. It also includes a PaaS layer, which is based on Pivotal CF, but which is not the official Pivotal CF. The following diagram shows the relationship between the PaaS services and the IaaS services in Helion.
HP’s Expectations for Helion
It is clear from the Helion presentation that HP believes that existing public clouds (Amazon, Google, Microsoft) are not enterprise grade and that this has limited the adoption of public cloud computing on the part of enterprises to five percent. HP believes that an enterprise-grade cloud platform like Helion will allow more enterprises to adopt public clouds, allowing adoption to rise to forty percent of enterprises by 2018. This is depicted in the following diagram.
Challenges Managing the Complexity of OpenStack
The next diagram depicts the different packages that comprise OpenStack. It would be safe to say that it would be a challenge for any one vendor to manage such a huge, complex, and multifaceted software product.
But OpenStack does not belong to just one vendor. It is, as we all know, an open-source project with many contributors. The major contributors to the Icehouse release of OpenStack are shown below. HP is behind Red Hat and IBM in the ranks of major vendors contributing code commits to Icehouse. What is really interesting is that Rackspace, the supposed ultimate champion of OpenStack, is behind Red Hat, IBM, and HP (certainly fertile ground for a follow-up post).
Is Helion a Viable Public Cloud Strategy?
The above brings us to the single most important question. Amazon, Microsoft, and Google have well-understood models for their public cloud businesses. Those models are all based on the vendor’s having absolute (or nearly absolute) technical control of its code stacks, and therefore having the unilateral ability to use Agile Development and DevOps to rapidly evolve its offerings. VMware discovered that charging cloud vendors license fees for its vCloud offering made the resulting offerings uncompetitive with Amazon. VMware has since launched its own hybrid cloud service in an attempt to take control of the pricing and positioning of its cloud offering with respect to Amazon. The approaches of the various public cloud vendors is shown in the table below:
|Cloud Vendor||Hypervisor||Cloud Management||Stack under Full Control of Cloud Vendor||License Fees Paid to Software Vendors||On-Premises Version Available||Quality of Service Level Guarantees|
|Amazon EC2||Owned by Amazon (derived from Xen)||Developed and owned by Amazon||Yes – Amazon has full control of all source code||No||Yes – but only through a separate company, Eucalyptus||Abysmal to non-existent|
|Microsoft Azure||Owned by Microsoft (derived from Hyper-V)||Developed and owned by Microsoft||Yes – Microsoft has full control of all source code|
Yes – over time as Azure components get integrated into Windows Server
At the application layer through partnerships with AppDynamics and New Relic – no end-to-end service level guarantees
|Google Compute Cloud||Red Hat KVM||Developed and owned by Google||No – Google relies on Red Hat for KVM but has full control of its cloud management layer||Business relationship between Google and Red Hat is unknown||No||Google has acquired StackDriver, which should lead to a solution in this area|
|VMware Hybrid Cloud Service||VMware ESXi||VMware vCloud Director||Yes – VMware controls its own hypervisor and its entire cloud stack||No – however, in some cases the VMware Hybrid Cloud Service runs in third- party data centers for which VMware has to pay||Yes – The VMware vCloud Suite is effectively the on- premises version of HCS||The VMware vCloud Suite includes extensive operations management and monitoring through vCenter Operations; no application level monitoring is provided|
|HP Helion||KVM and ESXi – Hyper-V coming||OpenStack plus HP additions||No – HP relies on the OpenStack open source community for a significant amount of the functionality in Helion||HP probably has to pay Red Hat for KVM licenses for its public cloud; customers running Helion on premises will have to buy their own hypervisor licenses||Yes||No mention of service quality monitoring so far|
|Virtustream||Choice of VMware ESXi or PowerVM||Virtustream xStream||No – Virtustream relies on either ESXi or PowerVM for the hypervisor but owns and controls its own cloud management layer||Yes – Virtustream has to pay for the hypervisors, but not for the cloud management layer||Yes – xStream S||Yes – Virtustream uniquely offers end-to-end application response time guarantees for business-critical applications like SAP|
The table above points out an obvious disparity between the vendors that have complete control of their software stack (Amazon, Microsoft, Google, and VMware) and vendors like HP, who are reliant upon third-party software for the core of their offerings. This disparity produces two effects on the competitiveness of the respective offerings:
- Amazon, Google, Microsoft, and VMware all own their entire software stacks and therefore do not have to pay license fees to any other vendor. This leads to these vendors having a cost advantage over any vendor that does have to pay license fees (which is why the third-party clouds that were based on vCloud all failed, leading VMware to launch its hybrid cloud service itself). HP might be at a minor cost advantage for its public cloud offering if it has to pay license feeds to Red Hat for KVM. For on-premises installations, HP could be at a significant disadvantage with respect to VMware and Microsoft, since neither vendor has to pay anyone else for their hypervisor.
- Perhaps more importantly, the question of economics (i.e., who is cheaper) is the question of who can bring enterprise-grade capabilities to the market quickly. This boils down to agility. And it raises the fundamental question that has to be asked with respect to the viability of any cloud offering based on OpenStack. It is highly unlikely that a consortium of IBM, Red Hat, Rackspace, and at least ten other contributing companies will be able to keep up with the teams at any of Amazon, Google, Microsoft, or VMware. The overhead (time wasted) in coordinating work between that many companies likely means that any public cloud offering based on OpenStack will always lag behind offerings from highly focused single vendors like Amazon, Google, VMware, and Microsoft.
- OpenStack also faces one more significant challenge. There are 10’s of millions of VM’s running on VMware vSphere on premise at customer sites using ESXi. There are 10’s of millions of VM’s running in Amazon’s public cloud using Amazon’s variant of the Xen hypervisor. OpenStack as a public cloud offering will almost always be deployed on KVM, which means that none of the existing workloads in the world of data center virtualization or public clouds will migrate easily to a KVM based cloud. If HP is serious about solving this problem they would need to establish a relationship with either Racemi or Hotlink.
What This Portends for HP
HP, like any company that primarily offers expensive hardware, software, and services, is being put under tremendous pressure by public cloud vendors who have turned hardware and software infrastructure into an affordable and easily consumable service. For HP to compete in the public cloud arena, it will have to transform many of its legacy, expensive-to-buy, expensive-to-implement, and services-heavy offerings into more affordable and consumable extensions to Helion. HP has stated that it intends to do precisely this. But in so doing, it will have to compete with Amazon and Google on price, which will put tremendous pressure on HP’s revenues. HP therefore faces the classic innovator’s dilemma. If it does nothing, then Amazon, Google, Microsoft, and VMware will erode its on-premises enterprise computing business over time. However, if HP succeeds with Helion, the resulting revenue stream will likely be less than the current revenue stream, based on high on-premises enterprise computing pricing. As such, HP faces the choice of watching its business get commoditized by others or participating in the commoditization of its business itself. HP therefore has to thread a needle. If HP fails with Helion, others will take HP’s business from HP. If HP succeeds with Helion, it will probably have to restructure the company in order to be profitable.
Helion is HP’s attempt to offer an enterprise-grade public, hybrid, and private cloud computing offering that can prevent customer defections to Amazon, Google, Microsoft, and VMware. Since Helion is based on OpenStack and there are no prior examples of successful public clouds based on OpenStack, HP faces some significant challenges.