On the twenty-first of October, HP announced that it is shutting down its Helion Public Cloud, which it built to compete head-to-head with AWS, GCS, and Azure. According to HP Cloud executive Bill Hilf, it is doing so to concentrate on helping its “customers to build and run the best cloud environment suited to their needs.”
This seems to make sense. Helion has often been under fire for being too small, too geographically focused, and too unfocused on what the public cloud actually needs from a public off-premises cloud. This is understandable. HP is a monolithic, enterprise-focused company, and its sales forces are laser focused on large, multimillion-dollar deals. Shutting Helion down will allow these folks to concentrate on what they are actually good at and use AWS for public cloud off-premises requirements. Well, as the popular idiom goes, “If you can’t beat them, join them.”
Now, this seemed to be the end of the news: a swan song of AWS beating another competitor into submission. However, fast-forward to October 29: at the OpenStack summit in Tokyo, HP announced a phoenix in the form of Helion 2.0. This apparent U-turn in fortune for the Helion product seems off. Didn’t we just say that Helion was too small to be a decent player in the high-stakes, low-margin public cloud business?
What makes this iteration different from Helion 1.0?
Version 2.0 will once again be a customized OpenStack cloud based on the previous OpenStack “Kilo” release. However, this time it will strictly adhere to the OpenStack API, allowing greater compatibility with other OpenStack deployments and the ever-growing number of verified third-party OpenStack plugins.
According to Bill Hilf, “Enterprises want to benefit from the powerful capabilities of OpenStack technology, but they must have the enterprise-grade capabilities required to support their businesses. The configuration, security and scalability advances in HP Helion OpenStack 2.0 enable organisations to deploy OpenStack technology into production with the confidence that they are backed by the experience and support of a trusted end-to-end technology partner.”
Other enhancements announced are that patch management is now continuous, so systems can be patched or upgraded with zero downtime on the infrastructure without breaking application availability. This obviously assumes your application is resilient. In fact, by using the built-in lifecycle management tools, an entire cloud can be upgraded without any outage.
Now, this release is one step behind OpenStack, as OpenStack has recently announced “Liberty,” its twelfth iteration of the product. However, as it is common for users of OpenStack in production to be conservative with regard to new releases, this is not considered an obstacle. In fact, Red Hat’s Enterprise Linux OpenStack Platform is also only running Kilo.
Some of the HP customization concerns hardening by the use of a hardened version of HP Linux and integration with other HP products like HP OneView. Helion also has a development platform based on Cloud Foundry to support customers on their journey to cloud-native apps.
Also of note, HP has introduced into Helion 2.0 the ability to create software-defined networks using HP’s Distributed Cloud Networking and Nuage Networks’ Virtualized Services Platform (VSP). This is an interesting play. Nuage VSP is an interesting platform offering software-defined networking (SDN) services to containers (Docker), hypervisors (KVM, vSphere, Xen), and virtual machines, as well as physical bare-metal resources. This will enable the provision of all resources to the cloud, including previously unavailable bare-metal resources. Now the vision of an all-encompassing private cloud is a possibility.
It is the integration of the SDN feature set that makes this new announcement all the more fascinating. This, to me, is not a public cloud play, but more of a hybrid and private on-premises cloud product. This fits more easily with HP’s sales engines: they can productize this and sell it to their large clients. With Helion installed as your cloud provider of choice on-premises, and with Helion’s adherence to OpenStack APIs, there is nothing to stop you from being able to burst into a public OpenStack-based cloud like Mirantis or even a yet-to-be-announced public Helion 2.0–based cloud. This version is much more compelling. The only thing that is confusing is why the company twilighted Helion 1.0 when all the while it was planning the release of Helion 2.0. Perhaps it has something to do with the left hand’s not knowing what the right hand is doing, what with the split that is occurring this month.