OpenStack Part II: OpenStack 2016 Survey Reveals Pattern of User Deployments
OpenStack use, and the pattern of users’ new deployments, were shown at the OpenStack 2016 conference in Austin, Texas, via the highlights of the OpenStack Foundation’s own survey of more than 1,600 respondents worldwide. The overview of survey results, accessible online via a Google search, casts more light on an important technology affecting the build-out of new cloud infrastructure.
The April, 2016, survey was the seventh one conducted by the OpenStack Foundation, to see how user adoption patterns are changing, over time. The shift in deployment types it reveals is significant, because new OpenStack releases are rolled out each year. This cycle roughly coincides with the two OpenStack conferences held annually.
The Year 2016 saw production rollouts of the Liberty release cycle, which followed the Kilo, June and Icehouse releases in 2015 and 2014. The next release is named Mitaka – with an emphasis on user experience, manageability and scalability. We point out that all of these capabilities are key focus points for enterprise deployments. Specifically, Mitaka is aimed at providing consistency for end-users; finer-grained controls and easier management for operators; and more scalability for larger deployments. (Please refer to Part I of my OpenStack Blog for more analysis of the overall OpenStack 2016 conference).
Key Findings for the April, 2016 OpenStack survey
The key findings for the survey were based on replies from 1,603 respondents worldwide. Of these, 26% were from North America, 24% were from Europe and 24% were from Asia/Pacific.
We think this tracks with overall IT usage worldwide, with much smaller percentages coming from South America, Africa and the Middle East – all areas with historically lower IT adoption, but where IT usage is growing rapidly today.
Key findings for the OpenStack survey were, as follows:
- OpenStack is most heavily used by large enterprises. The survey found that 16% were from organizations with 100,000 or more employees; 15% were from organizations with 10,000 to 99,999 employees and 23% were from organizations with 1,000 to 9,999 employees. The rest accounted for 11%.
- The average OpenStack cloud deployment runs 11 OpenStack projects, respondents said. This reflects gradual adoption of the OpenStack technology, surrounded by many more traditional enterprise deployments of applications and databases. This means that OpenStack is generally additive to existing enterprise workload deployments of SAP, Oracle and Microsoft. Even so, the delivery mechanism for those enterprise workloads may change, via the use of containers and virtual machines (VMs) across new layers of OpenStack cloud infrastructure.
- Many managers of OpenStack projects have multiple titles, by percent of respondents surveyed. The most typical roles include: cloud architect (43%); cloud operator or sys-admin (41%); OpenStack upstream developer (29%); application developer or deployer (30%); CIO or IT manager (14%) – and all others accounted for 15% of respondents. These findings reflect OpenStack’s highly technical audience; we think the supporters of OpenStack should come to include more business managers and LOB managers who see business value in OpenStack deployments.
Top OpenStack Deployments, by Project Type
Based on the OpenStack Foundation survey, the top OpenStack projects being used in production are the following:
- Nova compute service, with respondents reporting that 66% of Nova projects are in production, and 31% are in test mode.
- Keystone identity service, with 63% in production, and 30% in test mode.
- Glance image service, with 62% in production and 29% in test mode.
- Neutron networking service, with 57% in production, and 33% in test mode.
- Horizon dashboard service, with 59% in production, and 31% in test mode.
- Cinder block storage project with 57% in production, and 26% in test mode.
- Heat orchestration service, with 39% in production, ad 29% in test nmode
- Swift object storage service, with 32% in production, and 21% in test mode.
- The rest of the OpenStack projects were in earlier stages of deployment, with more user projects in the test mode and proof-of-concept mode than was the case for the projects listed above.
Looking at the survey results, it’s clear that large organizations have had the deepest experience working with OpenStack as they build out their cloud services. But it’s also true that large numbers of medium- and small-sized companies are working with OpenStack technology today. Developers at SMB firms are contributing to OpenStack projects, and are a significant part of the overall OpenStack community.
Often, OpenStack deployment case studies are cited from large organizations, with large OpenStack projects, such as: Wells Fargo, Visa, Volkswagen, and Walmart for private-cloud and hybrid cloud deployments; Deutsche Telekom, OVH.com and Go Daddy for public cloud; AT&T, China Mobile, NTT Docomo and Verizon for telecommunications; and the Texas Austin Computing Center (TACC), MIT, Clemson University and Los Alamos National Laboratory for scientific research.
The large enterprise organizations have the IT skill-sets and the IT budgets to fund applications development for leading-edge technologies such as OpenStack. They also like the idea of multi-vendor deployments with greater user consistency (e.g. consistent access and consistent app dev tools) via standard APIs.
It’s also true that smaller and medium organizations (SMB) are looking to OpenStack for cost reduction, and to gain improved interoperability in multi-vendor computing environments. Specifically, this will help the move to hybrid cloud – spanning enterprise and public-cloud computing. That is coming about not only through direct deployments, but also through partnerships with MSPs, system integrators, telecoms companies and Software-as-a- Service (SaaS) deployments.
There is a learning curve for OpenStack technology and for the community development projects that support them, as the survey results show. However, building cloud-enabling software through open-source community work and OpenStack technology is allowing the user base to broaden – and it’s supporting a wider pattern of adoption via a growing OpenStack ecosystem.