Sam Ramji has just been announced as the CEO of the Cloud Foundry Foundation. In addition, nine inaugural baord members have been added. The Cloud Foundry Foundation was formally established in December 2014 to advance Platform as a Service. With founding Platinum members EMC, HP, IBM, Intel, Pivotal, SAP and VMware in addition to more than 40 other companies, the foundation represents broad industry support for the project.
As a developer, being able to cf push Java and Scala applications with relative ease to multiple Cloud Foundry installations (a laptop running the open source version, Pivotal’s run.pivotal.io and IBM’s BlueMix in this case) qualifies as the tip of iceberg beyond the power of Cloud Foundry. The real power of Cloud Foundry is being able to integrate multiple services and let the platform handle the essential non-functional requirements such as scaling, high-availability, and security in a vendor-neutral fashion.
InfoQ caught up with some members of the foundation – James Watters, VP for Product at Pivotal, Christopher Ferris, CTO Open Cloud at IBM and Dr. Nic Williams, CEO of Stark and Wayne (which is a Silver member of the Foundation).
InfoQ: Does the formation of the Foundation affect developers at all? If so, how?
Watters: The foundation is a huge win for developers because today developer friendly, self service, cloud tools are often only available from providers like AWS. By rallying a broad coalition of enterprises users, systems integrators, service providers, and systems vendors the foundation is ensuring that developer centric infrastructure becomes ubiquitous.
InfoQ: Cloud Foundry clearly targets developers. However, developers are not the decision makers in many enterprises. The transformation of an IT shop to Cloud Foundry is not a trivial exercise. What is the plan to convince the CEOs and CTOs and even system administrators that moving to Cloud Foundry is the right move?
Watters: As you may have seen covered by Matt Asay in the Register, Pivotal Cloud Foundry made history with its first year sales results, exactly because Pivotal has earned so much trust from Sr. IT executives looking to transform their organizations. As the creators of Cloud Foundry, as well as Pivotal Labs, Pivotal really has come to be the brand most known for their forward thinking on development practices and platforms. In 15 years in enterprise tech, I’ve never seen enterprise leaders more focused on how to transform their organizations to be more cloud and developer centric.
InfoQ: The Foundation and the software has a largely Linux feel to it. Since Cloud Foundry is targeting enterprise development that has pockets of .NET development, how does the foundation plan to address that?
Watters: One of the harder decisions we made last year was when to add .NET support to our Elastic Runtime codebase. There has been amazing work done by the folks at CenturyLink on Iron Foundry, which extends the V2 Runtime environment to .NET. With the Diego refactoring of Elastic Runtime we had a great opportunity to make Windows and .NET first class citizens in the code without modification. We have now shipped alpha versions of that .NET support. Many of Pivotal’s customers are very interested in .NET and contacting both us and Microsoft for support of Windows, as well as Azure support within CF BOSH. I wouldn’t be surprised to see Microsoft join the Foundation in 2015.
InfoQ: Does the formation of Cloud Foundry Foundation have any parallels to the Java Community Process? Are there any lessons to be learned from that?
Ferris: Not really. We actually took quite a few cues from the Eclipse Foundation as we were designing the governance model for the Foundation. We wanted to preserve a degree of autonomy of the various projects, while at the same time ensuring that there was coherence of the whole. Thus, we have Project Management Committees (PMCs) and a PMC Council which provides overall architectural guidance and helps ensure cohesion between the various PMCs. We also wanted to give a voice to the users/operators, to provide a strong complement to the developers who drive innovation and contribute to the advancement of the various Cloud Foundry projects.
InfoQ: Interoperability between the Cloud Foundry installations will be very critical going forward. How does the Foundation plan to address this?
Ferris: Good question! We worked throughout the formative months of the Foundation on an approach that would allow the Foundation to get a solid head start on ensuring interoperability and portability of the various Cloud Foundry offerings, services and runtimes. We felt strongly that we needed to have this established at the outset. We have defined two certification programs: one for Cloud Foundry based offerings, which requires select components to be implemented in order to use the Powered by Cloud Foundry branding, and another for Services and Runtimes which intends to ensure that they can be integrated/used within a certified Cloud Foundry offering.
InfoQ: IBM is planning on integrating a number of services into their offering? Is this primarily how vendors will be able to differentiate?
Ferris: Yes, I believe that the greatest value is for a common platform shared by all. I believe that the real differentiation should be in the services that can be inter-operably/portably integrated with that common platform. That's where I see the greatest potential for vendors and consumers alike. Over time, I believe that the certification programs will evolve to allow certain differentiation of implementation by the definition of APIs and SPIs and corresponding tests that ensure interoperability of implementations, and there will be opportunity down the road for some to innovate in that regard. However, as I said, I think that the real value of an open platform such as Cloud Foundry is for there to be a high degree of interoperability amongst the various offerings, as this will help foster a vibrant ecosystem of portable applications, services and runtimes.
InfoQ: CloudFoundry is being touted as a panacea for the evolving style of devops and Microservice-based development. Any note of caution regarding this hype?
Dr. Nic: “Caution” for some, but “optimism” for others. Developer teams will finally have the ability to get operations tasks done by them and done quickly. Minimum-viable product (MVP) applications can be delivered to internal and external customers without explicitly involving the old school IT organization. New “app ops” teams may need to be formed. The responsibility for business success will sit squarely with the development/product team. The “caution” is for them to ignore the responsibility. The “optimism” is for those development/product teams who want to fail fast, want to setup modern continuous delivery pipelines, want to learn from their customers within weeks instead of years. Cloud Foundry is the great enabler for development/product teams.
Your team wants a complex continuous delivery pipeline combining micro services being developed from many other teams? Cloud Foundry is the orchestration platform that enables it. You choose the CI system - Jenkins, GoCD, etc.
InfoQ: You have been an active member of the Java and Ruby communities. Will Cloud Foundry help these developer communities specifically or developers at large?
Dr. Nic: Programming language communities have two axes of evolution - the ecosystem of libraries and tools to build better software faster; and the processes of decision-making, delivery and feedback from running applications. You can learn about new processes like Scrum or Agile or Continuous Delivery from your programming language community or local meetup group, but then discover your own company has so much IT concrete in place that you can’t evolve.
Cloud Foundry is exactly the platform that modern software developers need to support both axes of evolution. It supports all modern frameworks and programming languages and through buildpacks can be extended by any user to support anything that comes in the future. It also provides users (developers and app ops) with the command line tools, IDE integrations and APIs to evolve their internal delivery processes over time.
InfoQ: Many folks are complaining that the platform is already too bloated. Any thoughts?
Dr. Nic: Cloud Foundry is not your average open source project - those that are hacked together by part-timers in their spare time; or developed for one specific use case at a company and then “open sourced”. It has been carefully product managed by Pivotal for the last 5 years and actively developed by full-time developers and product managers.
It’s a large system of micro services - not a one monolithic Debian package service that many IT admins might expect. It’s designed to be run on just a few machines and to scale out to thousands of machines as your organization increasingly adopts it and the applications running upon it grow in size.
Operating a large system (which itself is responsible for orchestrating other applications/micro services) is not simple. Without help each IT organization would stumble towards working platforms eventually. Instead to simplify the initial deployment of Cloud Foundry - all the different micro services and internal data stores - there is Cloud Foundry BOSH. BOSH is also used to upgrade between weekly releases of Cloud Foundry, and to scale out Cloud Foundry to support more running applications over time,
Finally, BOSH heals the infrastructure being used to run Cloud Foundry. If you are using OpenStack and one of your OpenStack Nova nodes dies, BOSH will migrate the lost virtual machines to other Nova nodes. If AWS EC2 instances die, BOSH will provision new ones. Cloud Foundry is highly available itself. It will continue operation during the few minutes that underlying virtual machines go missing. Then a few minutes later BOSH recreates the missing infrastructure.
Cloud Foundry isn’t “bloated” in any sense I can think of. Perhaps it could actually do more - it could run Docker containers (coming in 2015).
Detailed documentation for installing, running and extending Cloud Foundry is in the docs.
