Cloud Foundry’s announcement of a new Container Management System is interesting…


Several years ago, Cloud Foundry (CF) created a multi-cloud PaaS platform and also made it open source. If my memories don’t fail me, CF was launched in response to Google App Engine, a hosted PaaS from Google, that was neither available for on-prem deployments nor as open source.


CF recently announced [1] that it is launching a new Container Management System (CMS) called Diego. CF has been deploying applications (that use their open source PaaS platform) in a container-like resource-constrained runtime environments. These application environments used CF’s proprietary application execution format called warden[2] although it was made open source [3]. The environments themselves are managed using CF’s Droplet Execution Agents (DEA)[4].

With this new announcement, CF is replacing its proprietary warden format and making it docker-compatible. Alongside they are also rebranding (and potentially upgrading) the management infrastructure to be Diego. Let us analyze this using the SWOT framework.

Strengths: This move is both significant and note worthy as it seeks to break the two-horse race led by Kubernetes [5] and Docker Swarm [6]. Moreover, CF’s experience in building a product and also servicing (demanding) enterprise customers puts it in a unique vantage point to take on the challengers.

Weaknesses: CF faces huge challengers — Google and Docker. Google is bringing its entire 15 years of experience in managing containers into Kubernetes while Docker seeks to provide an integrated orchestration platform. While some might argue a challenge by itself is not weakness per se, Google’s container experience definitely predates CF PaaS product timeline. In addition, as Docker keeps its product roadmap close-to-its-chest, CF might be at risk of always catching up with incumbent.

Opportunities: CF has a transient advantage here — its multi-cloud PaaS experience and its existing enterprise customers who are presumably using its products in an on-prem setting. CF will certainly work with its biggest partner IBM whose Bluemix [7] product is essentially enterprise-supported CF platform. In addition, given the open nature of today’s software development customers are likely to compare Diego with Kubernetes and Swarm using feature check table. Perhaps in addition to making Diego docker-compatible, CF Inc could also consider supporting a competing container format CoreOS’s rkt [8] which is lacking a CMS. Put together, Core OS + rkt + Diego is a good acquisition target as well :-)

Threats: By dumping its own container format and seeking to challenge Kubernetes, it appears as though CF Inc is taking on two battles at the same time — Docker Swarm and Google’s Kubernetes. Second, Docker being a startup and a very valuable one, is a good acquisition target and some recent rumors have indicated that Microsoft could be suitor which is in a acquisition mode for the last few years. If we assume this goes through, there is no risk of Docker being closed as the new Microsoft is more community-friendly and open. However, Microsoft is likely to put in significant resources to develop Swarm, essentially turning container ecosystem as a Microsoft Vs Google war. Another factor is consider is whether IBM will continue to use CF Inc’s Diego platform or will it switch to Swarm or Kubernetes.


Competition is always good in any industry and I personally like CF’s move both from strategic and survival perspectives. The latter is more important and timely for CF as the traditional cloud computing boundaries made of SaaS/PaaS/IaaS are blurring. Only time will tell if there is a sufficient market size for a 3rd CMS.


One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.