Is OpenStack dead?Paul Adam
Last Updated: July 2018
There is plenty of news around on OpenStack than about anything else put together in the world of IT over the last few months. Latest news was PLUMgrid, a local bay area startup, which went under resulting in IPR fire sale to VMware.
Let me summarize my understanding and what is happening as a whole here in a few threads:
1. Legacy ICT vendors and Public Cloud: AWS continues to claim victims, the most recent ones were VMware, HPE and Cisco who all abandoned their public cloud initiatives. This was long coming, HPE/Cisco were using OpenStack as their public cloud platform, so their departure reflects negative on OpenStack. VMware essentially put vCloud Air to rest!
AWS is on a tear with revenues now at $13B/year run rate and continues to grow at over 30% CAGR.
2. Mirantis and remaining pureplay OpenStack startups: OpenStack vendor M&A activities already happened during 2015 where HPE, Cisco, EMC all picked their winners leading to market consolidation in that space. Mirantis is continuously pivoting to continue to be alive, however long term they will either be acquired or run out of funding. PLUMgrid had been looking for a buyer for some time, and eventually liquidated. OpenStack, just like all other open source, is and will be difficult space for startups given their cash constraints and limit on pivots. I don’t expect to any significant venture funding going into OpenStack.
3. RedHat, SUSE and Linux Foundation: With all the news around, it does seem that Linux crowd will be the new custodians of OpenStack and busy acquiring related assets. OpenStack promotes Linux (as opposed to windows), so that is a good match. RedHat also the only company that knows how to build profitable business using Open source (Horton Works is another one, but in big data space).
4. Private Cloud and OpenStack: It is an area to watch, as AWS continues to roll ahead, there will be plenty of news on private cloud including Microsoft long-awaited Azure Stack slated to be launched in mid-2017 with HPE and Dell as HW infra partners. AWS has proven the public cloud economics, availability, reliability, scale, features are all sufficient and could never be recreated and matched by any customer in-house. All of this means bad for private cloud, however I do believe that managed private cloud and esp. in niche GEOs will remain around. However, I do believe, far more spending will take place in public cloud versus private cloud as we move forward. There are still reasons for private cloud such as regulatory and compliance, so companies in regulated industries will continue to need highly controlled build-outs.
In a nutshell, I don’t believe OpenStack is dead but it has definitely gone through near-death experiences with plenty of bad news in the past in the form of mass corporate exodus (due to AWS, not as such due to OpenStack), private businesses realizing they could never rebuild an AWS-like cloud in-house with anything (including OpenStack), and last wave of startups in OpenStack space going under. Walmart recently announced that they are moving from AWS to Azure and will keep their private OpenStack cloud for now. Going all-in to public cloud is simply not going to work for a large enterprise like Walmart, if you’ve ever worked in an Enterprise, you know they are slow, loth and heavily process oriented.
However, there is hope if you see what’s happened at the recent OpenStack Summit (Vancouver May 21-24 2018). Let me summarize a few key points from this summit.
- Focus seems to be shifting to ease of deployment around OpenStack with two potential candidates that may be included into either OpenStack Foundation or CNCF. Those CI/CD candidate are Spinnaker (from NetFlix and Google) and Zuul (from OpenStack Foundation).
- There is no doubt that Kubernetes and CNCF/Linux Foundation continue to move ahead and OpenStack Foundation clearly sees that opening up to cloud native development model and thus newer projects is essential. OpenStack Foundation seems open to the idea of running Kubernetes on top of OpenStack.
- It is clear that OpenStack Foundation needs to get over the “container as a threat” idea and do what’s right, which is to embrace containers and make deploying infrastructure easier. Virtually every customer is running VMware and wants to transition to containers with Kubernetes orchestration as fast as they can.
- OpenStack Foundation doesn’t seem to have all the answers, so we shall see if it is Kubernetes on bare metal or Kubernetes on OpenStack or some combination where Kubernetes can make use of only core OpenStack services.
- Last but no least, China investment seems trending up into OpenStack Foundation, and recently Tencent swapped out IBM as a member.