I hear from almost every Enterprise and service provider that vCloud is not what they want. Why? A significant concern is that it locks you into a single vendor cloud model that is utterly undifferentiated yet appallingly expensive, nickel and dime-ing you on a per-VM basis for every possible feature. Project Redwood – a massive undertaking from my friend the Sheriff of Redwood Valley to build the perfect, yet perfectly mismatched cloud platform – really is deadwood.
We in the OpenStack community believe that you ought to have a choice. We don’t think you need to throw out ESX or even vSphere. You made a rational decision to use (err, buy) it. But you ought to have a choice as to whether your cloud implementation will lock you into a single vendor model forever, with a limited set of expensive value-added services. We think there’s another way, that permits innovation in cloud services and solutions, that scales massively, and that is wholly free. The answer is OpenStack Compute – a massively scalable cloud orchestration system developed by over 50 vendors, hundreds of engineers and put in production by the world’s largest cloud service providers. Just to put this in perspective, here are some recent stats from the OpenStack community:
- Data Tracking Graphs
- OpenStack Compute (NOVA)
- 22 Active Reviews
- 192 Active Branches – owned by 48 people & 10 teams
- 2,320 commits by 63 people in last month
- OpenStack Object Storage (SWIFT)
- 4 Active Reviews
- 47 Active Branches – owned by 19 people & 4 teams
- 189 commits by 13 people in last month
And now NOVA supports vSphere. The hypervisor-vmware-vsphere-support blueprint merged yesterday into mainline, so Nova (OpenStack’s compute cloud) can now use VMware’s hypervisor for compute nodes. This code has been extensively peer reviewed and tested. This brings the OpenStack compute options to seven: vSphere, XenServer/Xen Cloud Platform, Xen, Hyper-V, KVM, QEMU, and UML. LXC isn’t far behind, which will make eight. Still wondering why you paid for an expensive-yet-commoditized hypervisor?
We don’t normally do Oscar ceremonies for branch merges, but this warrants a huge shout-out to the key folk from the OpenStack community who were involved. It is also worth noting how quickly this project was accomplished (3 months to production ready code) relative to the traditional enterprise 2-3 year product cycle (Redwood trees take thousands of years to grow):
- Ewan Mellor, OpenStack Architecture Board member, and Chief Architect for Citrix OpenCloud
- Ravi Gururaj and Milind Barve for project requirements & management
- Neeraj Koul, Rochanglien Infimate and Sateesh Chodapuneedi for ESX support and for automated QA.
- Armando Migliaccio, Salvatore Orlando, Jay Pipes, Rick Harris and Rick Clark for extensive code quality peer review
- Anne Gentle for documentation
- Søren Hansen for updating the Hudson environment at short notice to get the merge completed.
- Thierry Carrez for release management
The commitment from the OpenStack community is to deliver you not only an alternative to vCloud, but the only possible choice if you adopt a rational approach to building cloud. You rationally chose ESX and vSphere. If you’re rational, you’ll choose OpenStack.