Rishidot Research

deciphering the modern enterprise

cloudfoundry

Docker Announces Support For Kubernetes – An Analysis

· · Leave a Comment

Yesterday at Dockercon EU, Docker announced its support for Kubernetes on the Docker Enterprise Edition, Docker Community Edition as well as its desktop apps as well as the Moby project. This is a significant shift for a company that almost broke the open source community around the then Docker project. They wanted to push the hooks for their orchestration and management plane into the containers under the “batteries included but swappable” marketing campaign. Since then, the wind has blown in the direction of Kubernetes at the orchestration level and the conversation has effectively moved from the standardization around containers to standardization on orchestration plane. In this post, we will discuss the implication of this announcement in the market and how it impacts IT decision makers.

Docker’s foray into Kubernetes World

Yesterday Docker pre-announced the availability of Kubernetes on Docker platforms and the Moby project citing the shared roots between Docker community and Kubernetes community. They also announced that they would make vanilla Kubernetes available and stay close to the recent version instead of the Red Hat model of releasing stable releases for OpenShift Container Platform. According to Docker, there will be better collaboration between the Moby project and Kubernetes project. The end users get the option of selecting Kubernetes or Swarm for orchestration.

The State Of Developer Platforms

It is all about application platforms. How do you empower developers in your organization to seamlessly deploy apps ensuring faster time to market? How organizations enable them depends on the abstraction which, in turn, depends on the nature and requirements of the application being deployed. The early days of cloud saw the debates of IaaS+ vs PaaS and we see similar trends in the era of container native workloads. Kubernetes is fast gaining mindshare, driven by the declarative approach it offers in the automation of container native infrastructure. The quest to pick the right abstraction needed for various applications still see the same kind of demarcation we saw in the early days of cloud computing. It is IaaS+ (driven mainly by Kubernetes even though Mesosphere DCOS and Docker Swarm are other competing platforms) vs the platform abstraction at the developer layer enabled by platforms like OpenShift and Pivotal CloudFoundry (picking Pivotal CloudFoundry specifically because I don’t see any other credible vendor in that ecosystem) vs the serverless or Functions as a Service offerings. The usage patterns range from monolithic and web apps in IaaS+ to Modern apps including Microservices on developers focussed platforms like OpenShift and CloudFoundry to event-driven Microservices in the Serverless/FaaS platforms.

The announcement by CloudFoundry that Kubernetes will become the Container Runtime for CloudFoundry platform combined with Docker’s announcement that Kubernetes will be one of the choices in orchestration plane puts Kubernetes as the core component in the container native application platforms. Kubernetes, by itself, has limited impact but it is emerging as the core component of modern day platforms whether it is IaaS+ or modern PaaS or FaaS. Both Pivotal CloudFoundry and Docker are positioning their support for Kubernetes as giving a choice to their customers. While this may be true in the short term, there is a high chance that Kubernetes will emerge as a standard in the container orchestration and be a standard component of any developer-centric platform.

In that sense, Kubernetes is fast emerging as a standard for container orchestration. But, we want to discount any notion that Kubernetes has won the platform wars. The platform market is wide open with many of the workloads still in VM machines and Kubernetes adoption in production is still in early stages. Functions as a Service (as a public cloud service) or a FaaS Platform that is multi-cloud and agnostic of orchestration layer may take the steam out of Kubernetes just like how Kubernetes took the winds off Docker momentum.

Considerations for IT Decision Makers

This makes the decision much easier for IT decision makers and it helps them consolidate their platform choices without worrying about whether the platform supports Kubernetes or not. If your organization has already invested in Docker Platform, this makes it easy to have a mixed environment where Kubernetes can be used for managing dev and test clusters and Docker Swarm for production. The next version of Docker Enterprise Edition and Docker Community Edition will make this easier for your organization. If you are not a Docker shop and want to have a choice in the container orchestration, it makes sense to go with Docker Platform. Otherwise, there are other choices from established vendors like Red Hat OpenShift or Pivotal’s CloudFoundry Platform. Between Red Hat OpenShift and Pivotal CloudFoundry, the decision is mostly cultural. If you are an IT-centric organization, Red Hat OpenShift Container Platform is well suited for your needs. If you are a developer focussed organization, Red Hat’s OpenShift Online or OpenShift Dedicated or Pivotal’s CloudFoundry are better options. Depending on the tolerance level of the organization for betting on startups, there are other options like Mesosphere DCOS, Rancher Labs, Heptio and many others. But if your end goal is to embrace Functions as a Service, you could still use containers to encapsulate the backend services but we would strongly recommend that you bet on multi-cloud, container orchestration agnostic platforms. It doesn’t make sense to embrace Kubernetes just for using FaaS.

Conclusion

Docker’s move into Kubernetes is the next logical step for them after they failed to capitalize on the momentum behind their container mindshare. This also makes them a much easier acquisition target as every big company has bet their modern stack strategy on Kubernetes. It will be interesting to see where Docker goes from here as Steve Singh takes full control with the newer round of funding expected to happen soon.

Briefing Notes: Cloudsoft AMP

· · 1 Comment

This is a briefing note prepared by me on Cloudsoft AMP, a DevOps platform services player offering autonomic management of applications.

Overview:

Cloudsoft AMP is an enterprise application management platform that helps automate the process based on business and performance needs. Deploying an application on any cloud or PaaS is just one part of the application lifecycle management. There are many other aspects of the lifecycle that are equally important. For example, management, monitoring, governance, portability, etc. play a critical role in enterprise IT. Cloudsoft AMP adds a layer of abstraction to platforms/platform services making autonomic management of applications much easier in enterprise IT environments.

Read the report below

Disclosure: Cloudsoft Corporation is a client and sponsor of Deploycon
To download the briefing note, you need to sign up as a free subscriber. Check out this page for signing up as a free subscriber. Once you sign up for your account and log in, you will see a download link to the briefing note.

[s2If is_user_logged_in()]
Download Link: https://rishidotver3.wpengine.com/?s2member_file_download=RishidotResearchBriefingNotesCloudsoftAMP.pdf
[/s2If]

PaaS Pivot: Big Data At The Core Of Platform Services

· · Leave a Comment

As we go into 2013, I keep thinking about the evolution of the Platform as a Service and wonder what is in store for this segment this year. As Platform Services are one of my core focus areas of research, I thought I will start off this year with a post on this topic. For the past year or so, I have been advocating the need to rethink PaaS offerings in order to fully take advantage of the big data paradigm. I use the term Intelligent Platforms to describe next generation platform services built around big data. In my opinion, we are going to see a pivot in the PaaS market where the focus will shift from the application development platforms focussed on scaling users and meeting the resource demands of large loads to building a robust platform to take advantage of vast amounts of data organizations have or going to acquire in the future.

Intelligent Apps Ver 1

Historically, platform as a service offerings were focussed on modern web applications that handle “smaller quantities” of data. In some cases, PaaS was used for applications that handle (or make use of) large volumes of data. But, in my opinion, most of these applications on these platforms were just scratching the surface. In reality, the PaaS offerings were not built to meet the needs of data hungry organizations. In 2012, PaaS vendors understood the changing needs of enterprises and were slowly starting to focus on the big data use cases. At the same time, we also saw the emergence of big data applications which were built on top of big data infrastructure platforms like Hadoop. As enterprises understand the full impact of big data and start building apps to take advantage of data (collected across the length and breadth of these organizations), they will really feel the need for more sophisticated platform services that run on top of big data infrastructure.

Vendors like Continuuity are trying to attack this problem. VMware’s CloudFoundry spinoff seems to be heading in this direction. In 2013, we are going to see the emergence of more such players and we will also see most of the existing PaaS vendor take steps to boost their platform services so that they are capable of supporting big data applications. Keep in mind that whatever we have seen so far with regards to “big data applications” are mostly focussed on analytics and visualization. What we are going to see in the future are set of services built on top of Intelligent Platforms that will go beyond simple analytics. We are going to see applications (services) that are self evolving and which can tweak itself based on the insights gleaned from various data sources (including the data these applications themselves produce). The underlying platform services needed to support such sophisticated services are going to be much more complex underneath than what we are seeing among the PaaS vendors today.

In short, 2013 will be the year when platform vendors are moving towards building platform services suitable for intelligent self evolving applications (services) of the future. All these services are going to be centered around data. Not just business and governments but the entire human society is going to rely on the data driven services with unprecedented complexity and automation underneath. The key for any vendor in the space is to build Intelligent Platforms that mask all these complexities and offer a simple interface for developers.

PS: The image in the blog post was a simplistic diagram I put forward to highlight the evolution from the current generation of PaaS to the next gen Intelligent Platforms.

Subscribe to Modern Enterprise Newsletter & get notified about our research