Hello, readers! This is Steve Henning and I'm the CMO of appLariat. I'll be writing about IT market dynamics and how they relate to adoption of containerization. I'll probably throw in some blogs on marketing specific topics as well, but I digress...
I was recently doing some research on the Hyperconverged Infrastructure Integrated Systems (HCIS) market and came across a very interesting article:
I certainly agree with the author, Chris Mellor that, at this point, the market is a two horse race between Dell/EMC and Nutanix, but I’ve got a different take on how HPE, Cisco (and others) can catch the two leaders.
I believe that making up ground in this lucrative market (Gartner says the HCIS market will be worth $5B by 2019) is not going to be about the best hardware or infrastructure management software. It’s going to be about how fast organizations (or their customers in the case of cloud providers and MSPs) consuming HCIS can move applications and application development teams onto these platforms in a way that speeds their time to market and increases their competitiveness. Prior to appLariat, I was the CMO at ServiceMesh ( a first generation Cloud Management Platform or CMP provider) and to illustrate, let me share the story of how we started working with Visa. Visa had made a substantial investment in VBlock converged infrastructure to increase competitiveness in the point of sale market and after nine months of working with VMware software and EMC consultants, they had not moved a single development team or application onto VBlock. EMC turned to ServiceMesh to provide a solution to move the apps and development teams to the VBlocks by turning the vSphere environments running on them into VM-based clouds. Initial POC success with a small development team led to ServiceMesh winning the business and in less than two months Visa had moved the team to the VBlock, something that couldn’t be done in the previous 9 months without ServiceMesh. Unfortunately, limitations of the VM-based application blueprinting inherent in ServiceMesh’s product (and all other first generation CMPs) led to less success with more complex applications. But clearly, the initial reason for the purchase of the VBlocks was to provide a platform for the dev teams to speed time to market for new applications and services.
Fast forward to dockerCon17 and Visa’s Swamy Kocherlakota, their SVP, Global Head of Infrastructure and Operations discussed how they were using container technology to make their electronic payment platform accessible to everyone, everywhere. Container technology had reduced application provisioning time and significantly improved their ability to maintain and patch. They are planning to move as many workloads to containers as possible to improve efficiency. This is not just about efficiency either. Apps that are moved to containers running on a modern container runtime like Kubernetes gain the cloud native capabilities of portability, elasticity and resilience without the need to change one line of code. One of the most interesting things Kocherlakota discussed was that, after careful analysis, they decided to run their container cloud on bare metal because it was clearly the most efficient approach. That certainly spells trouble for the future of the hypervisor.
And it’s not just Visa. The container market is already a $1B+ market that is growing at 40% year over year. Due to the efficiency gains and new capabilities associated with developing and delivering applications on containers, there is no question that the future delivery method for all applications will be container-based.
Now getting back to the HCIS market, I believe that in order for a clear winner to emerge, the vendor’s focus has to shift to delivering HCIS-based container clouds on bare metal because this will be the fastest way for development teams and IT Ops to consume them to deliver apps to address new market opportunities. There is just no reason for a hypervisor with containers, and removing that abstraction reduces cost to the vendor and the end customer. Of course, in order to achieve this, you need a container runtime to run on the HCIS. Fortunately, the best and most popular one available is Kubernetes, which is an open source solution. And, of course, you need a Container Automation Platform that
- allows existing applications to be moved to containers quickly without a massive learning curve
- deploys and manages Kubernetes and provides policy driven cluster/app scaling and other Day 2 operations
It simply isn’t enough to support new container and microservices-based development to win in the HCIS market. Large enterprises have hundreds, even thousands of existing/legacy applications in ongoing development and they are looking to move them to containers, a message that was very clear from those presenting at dockercon17. And you cannot expect every one of these dev teams to be trained to deliver apps in containers manually.
Of course, some of the HCIS players have acquired first generation CMPs and the tendency will be to try and bolt on container automation and management to their existing offerings to address this change in market dynamics. But anyone that has developed deep expertise in containers and runtimes like Kubernetes knows that this technology requires a completely new management paradigm and that the previous generation VM-based blueprinting mechanisms simply don’t apply. I believe the winner in the HCIS market is going to realize this and take a different approach, but delivering HCIS as an elastic container cloud is key.
Nutanix, in my opinion, made an extremely smart acquisition with Calm.io, which is a cloud management platform designed from the core on containers. While, it has somewhat disappeared into their software portfolio at this time, I believe they will soon realize what they have and they could become the clear winner in this market. It’s up to the other vendors in the HCIS space to decide if they want to let that happen.