How long before containers become the new virtualisation layer?

Let’s get something very clear from the start, Docker isn’t a replacement for application virtualisation. Instead it is a new, more agile, way of developing and deploying applications into production. Now I have that off my chest we can talk about how the world is moving from virtualisation to containers, in the same way that they moved from physical to virtual all those years ago.

There are several people in the market, primarily working around VMWare that are trying to convince people that you will need both moving forward, and that they are the best positioned to deliver that functionality. Whilst that may be true today, where organisations are using containers for niche projects and are still primarily virtualised, as that swing changes so will the requirement for the management and orchestration platform. The main question then becomes why would you pay a premium for a VMWare solution to manage a mainly containerised environment?

Based on this soon arriving new model, VMWare believes they will become the orchestration and management platform of choice for multi-cloud, multi-service based deployments. The question I have is why? More and more customers we talk to now either use multi-hypervisors or considering using multi-hypervisors, they have some form of cloud deployments (adding an additional management layer) and are using containers in some way. VMWare has always been the Oracle of the hypervisor world, in that they were reassuringly expensive, but for a long time the only game in town. That isn’t the case anymore, so why would you pay a VMWare premium if they are no longer the mainstay of the environment you are managing?

If you look at most of the organisations that are running our digital world today you would be hard pushed to find a VM. Companies like Google, Twitter, Facebook and the vast majority of the web based companies have already made the move to 90%+ container based application deployments. In a recent survey by Docker, 90% of companies interviewed are using them in application development and 58% of companies are using Docker in production. So even just by playing those numbers through over the next few years both numbers will grow substantially as the applications currently in development more into production.

The main problem with containers right now are the enterprise feature sets that are missing, around orchestration and management. As everything is included in the container; compute, networking, storage and control plane management across the entire environment becomes complicated and cumbersome as you could end up having thousands of containers running different microservices. Also, connecting all the microservices together isn’t the easiest thing in the world either…. Currently! With things such as Docker DC, CoreOS, Kubernetes, Openshift, Mesosphere and others making huge strides in providing the enterprise feature sets that organisations are looking for it is only a matter of time before this changes.

In the same way that organisation now run mostly virtualised with some physical servers for specific applications, I believe the future will be organisations running mainly containers with some virtual and some physical for specific applications that don’t work very well in containers. My prediction is that within 18 months you will see larger organisation primarily using containers as their application deployment model instead of VM’s, and within 3 years it will become the mainstream way that organisations of all sizes, deploy and manage applications. Will VMWare still be the mainstay of the application orchestration and deployment market? Maybe, but its gonna be fun to see how they keep themselves relevant with all the changes coming and the more agile, dynamic players in the space trying to steal their lunch.

«
»

Webinar

To watch this webinar please enter your details below.

Datasheet

To download this form please enter the details below.

The download will open in a new window, ensure that you site allows pop-ups.