Reflections on the PaaS market and Cloud Foundry Summit 2016
This week I spent a few days in sunny California for the 3rd annual Cloud Foundry Summit (US). The Summit, held May 23rd-25th in the Santa Clara Convention Center drew nearly 1,700 attendees, representing an increase of 13% over last year’s 1,500.
The event presented an opportunity to reflect on the state of the Cloud Foundry ecosystem as well as the PaaS market in general. Here are my key observations and takeaways on those topics.
1. Large Enterprise PaaS Adoption is Finally Here
I’ve been a PaaS cheerleader for as long as platform has been sandwiched between app and infrastructure in the Cloud Pyramid, so it’s been exciting to see the mainstream enterprise PaaS market opportunity finally start to materialize.
Some of the largest enterprises in the world have adopted platform-as-a-service as the basis for their digital transformation efforts, and the pace seems to be accelerating. Cloud Foundry users include GE, Ford, Volkswagen, and Humana and more–with the commitment on the part of GE and Ford being so strong as to drive an investment by those companies into Cloud Foundry vendor Pivotal.
The program at CF Summit reflected this growing enterprise adoption, and included speakers from Comcast, GE, GAP, Ford, Allstate, Orange, Kaiser Permanente, BNY Mellon, Visa, Thomson Reuters, Swisscom, the Government of Australia and more. While we’ve heard from many of these companies at previous US Summits, the list also includes a healthy number of new names as well: GAP, Kaiser Permanente, Springer Nature, and the Australian Digital Transformation Office are notable examples. GAP, for example, presented on a point analytics application that they migrated from AWS over to CF to increase agility and reduce cost.
Those that presented previously seemed to be making good progress as well. In its day 1 keynote, Comcast, which went into production with Cloud Foundry at the end of last year, described some of its results: Driving the time from an idea to a new feature down from weeks to 2-3 days, and the time to scale an existing production application from months down to minutes. And Allstate, which is still driving towards broad adoption, shared some of the insights they’ve gained into of the cultural adoption issues they’ve run into while driving adoption in a large, conservative enterprise.
All in all, the Cloud Foundry Foundation demonstrated strong evidence of growing customer adoption and success here at the Summit.
2. A healthy ecosystem has formed around Cloud Foundry
The Cloud Foundry ecosystem is growing quickly as well. The Foundation currently has 63 members and is growing at a rate of about one new member every couple of weeks, based, it was emphasized, all on inbound interest. Membership includes services/distro providers and complementary technology vendors, as well as end-users. HP, IBM and Pivotal all had high-profile customer case studies presented during the event keynotes, suggesting a playing field in which the various member companies all have an opportunity to benefit from Cloud Foundry’s growing acceptance.
The Cloud Foundry project itself saw a 36% increase in code commits from a base of over 2,100 contributors, with the percentage of commits by founding sponsor Pivotal dropping to about 65% based on growing contributions by other companies. The project has over 130 full-time dedicated committers from across the various member companies. Full-time committers are uniquely encouraged by the Foundation’s contribution model.
End-user interest in the project is strong as well, with nearly 175 independently organized local groups bringing together over 33,400 members in 105 cities world-wide.
3. The PaaS Wars are Over…
Following a wave of consolidation in the private cloud PaaS market (~2010-2012), the subsequent few years (~2012-2015) marked a period of intense competition among the remaining players in the market. Those players, principally Pivotal with Cloud Foundry, Red Hat with OpenShift, and Apprenda with its eponymous product, duked it out both in individual deals and in public skirmishes. We can now confirm definitively (if we couldn’t this time last year) that Pivotal delivered the master stroke in February of 2014 with the formation of the Cloud Foundry Foundation and the adoption of an open governance model for the project.
Cloud Foundry Foundation membership. pic.twitter.com/xU5dLvLcqK— Sam Charrington (@samcharrington) May 26, 2016
Things may have been different if Red Hat had gotten there first. Rumor has it that in early 2014 they were hoping to launch a foundation of their own. They have in OpenShift Enterprise 3 an arguably more open and technically superior product, with fundamental, architectural support for Docker and Kubernetes while the Cloud Foundry still struggles, clumsily at times, with its relationship to the broader container ecosystem. And they’ve got some great customer wins, with companies like Amadeus, and more recently BBVA, adopting the software to support broad digital transformation efforts.
In the wake of the success of Cloud Foundry and the Foundation, barring further technologic disruption, we’ll likely see the Apprenda (whose marquee customer J.P. Morgan Chase is a Silver Member of the Foundation) and Red Hat products relegated to niche roles for customers in the .NET and Red Hat ecosystems respectively.
4. Let the PaaS Wars Begin!
Did someone say disruption? The threat to Cloud Foundry’s all-in-one approach to PaaS has come in the form of container orchestration tools, namely Kubernetes. The point here is not that CF and Kubernetes are interchangeable, but rather that, should enterprises begin to adopt Kubenetes in a big way, they will want a PaaS that works on top of it, not as a separate silo.
We can see this starting to happen in the OpenStack community, where between 2015 and 2016 Cloud Foundry lost ground to Kubernetes in a big way. In the OpenStack Foundation’s annual user survey, production and dev/QA use of Cloud Foundry and Kubernetes in that community were equal in 2015 at 32% of respondents (n=82). In 2016, Kubernetes adoption rose to 35% of respondents (n=118), while Cloud Foundry use fell precipitously by 40% to 19% of respondents. And anecdotally, I’ve moderated a use cases track at the last couple of DockerCons and several presentations in each session have been engineering teams using container technology to build custom PaaS or PaaS-like systems. PaaS and container orchestration are certainly on a converging course.
In what’s shaping up to be a classic innovator’s dilemma case study, the Cloud Foundry community has resisted ceding ground to Kubernetes, rather promoting the virtues of its own container orchestration system Diego, while those players with less to lose have moved more quickly to support the new approach. Red Hat, to its credit, saw this shift early and the latest version of OpenShift was rewritten from the ground up around Kubernetes. Apprenda has pivoted towards Kubernetes more recently.
These new PaaS Wars will play out over the course of the next couple of years, even while the container orchestration vendors themselves duke it out.
5. Missing in Action: Technology Vision
One troubling omission from the keynote sessions at the CF Summit was any kind of technical vision on the part of the Foundation or CF community. The foundation spoke extensively, and quite eloquently, on the topic of community and ecosystem, we heard from vendors and customers on their progress, and we learned about robots taking over the world, black holes, and how we’ll soon connect our RasberryPis to our brains but nothing at all about the future of PaaS or Cloud Foundry from a tech perspective.
The Cloud Foundry community needs strong technical leadership and a guiding technical vision, particularly in light of the
coming second phase of PaaS wars as described above. The lack of any attempt at presenting a technical direction suggests either that stagnation or complacency have set in, that there’s no consensus around what the future needs to look like, or that that future is being decided among a small group of players behind closed doors and without regard for broader community engagement. None of these are promising signs.
Whatever you think about OpenStack and the OpenStack Foundation, this is an area they do a much better job in: Jonathan Bryce usually leads a discussion on the community and ecosystem story in the day 1 keynotes with Mark Collier orchestrating the sharing of the technology vision in the day 2 keynotes. (Though I felt this year’s tech vision discussion lacked some coherency, largely due to the inclusion of the Stackernetes demo with no broader contextual discussion.)
Hopefully we’ll see a stronger showing in this area in future CF Summits.
More on the Summit
For more thoughts from this year’s Cloud Foundry Summit keynotes as well as the Analyst Breakfast I attended, take a look at my blow-by-blow from Twitter, below:
Day 1: Keynotes
MyPoV: @MFordFuture's keynote raises important if sobering issues that we need to be thinking about and working on.— Sam Charrington (@samcharrington) May 23, 2016
Day 2: Analyst Breakfast & Keynotes
#2 Grow #cloudfoundry community talent base. Quickly.— Sam Charrington (@samcharrington) May 24, 2016
#3 There's a gap in training/enablement for product owners. What happens when ready to push to prod on #CloudFoundry?— Sam Charrington (@samcharrington) May 24, 2016
#4 Need to grow the ecosystem/community. Make sure they understand the implications of #cloudnative approach.— Sam Charrington (@samcharrington) May 24, 2016
#5 #cloudfoundry community needs to continue to up their security game.— Sam Charrington (@samcharrington) May 24, 2016
Summary of #cloudfoundry customer panel Top Asks: #1 Multi-cloud, #2 Talent, #3 Prod Owner enablement, #4 Grow community, #5 Security.— Sam Charrington (@samcharrington) May 24, 2016
Broader than provider, they need cross region even w single provider. https://t.co/cjOWtoXdeS— Sam Charrington (@samcharrington) May 24, 2016