Over its lifetime, Mirantis has proved to be adept at changing course to maximize its chances of success. While “pivoting” is a long-appreciated Silicon Valley dark art, Mirantis has taken the idea to even greater heights and its frequent pivots have led to the classification of an entirely new disorder, the whiplash-like “Mirantis neck.”
Joking aside, Mirantis has done a masterful job of skating, if not to where the puck is going, at least away from where the puck is no longer. First founded as a pure-play OpenStack company, Mirantis spent years building scale and momentum in this area and outlasted the other big vendors who had much more cash to invest in their OpenStack programs.
After failing to really achieve its originally stated outcomes, however, OpenStack has lost some of its luster and, importantly for Mirantis, seems less of a place to build a viable product business – something Mirantis is desperately trying to do. As such, the company made another shift a year or two ago, culling a large proportion of its team, and refocusing in the Kubernetes space. While still supporting OpenStack, it is fairly obvious where its future lies.
The Mirantis Cloud Platform is kind of product, kind of services play that brings Kubernetes and OpenStack together in order to help large enterprises build their own private clouds. And since infrastructure may not be as sexy as it once was, Mirantis also offers an Application Platform designed to help those same enterprises adopt new ways of developing applications – bundling up all the buzzwordy “continuous delivery” and “cloud native,” MAP is more at the “developer agility” end of the agile continuum.
The edge is the coolest place in town
Over recent years an interesting trend has occurred. As the cloud becomes a default, and every organization under the sun accepts the agility dogma, one of the new hot areas (apart from, of course, AI) is edge computing. The rationale for needing an edge play actually makes sense. The rise of the cloud has seen an ever greater proportion of the aggregate computing requirement focused into a small number of mega-scale clouds.
This trend has, somewhat ironically, reversed the distributed computing trend that the PC created, and recreated (at least in some ways) a far more concentrated computing paradigm, not dissimilar to the mainframe age. The issue with this concentration, no matter how efficient it may be, is that it ignores the very real issues of physics. By definition, these few mega clouds cannot process where the individual users actually are and hence distance becomes a factor.
Which is where the edge comes in. The idea being that applications can leverage these big central clouds for bulk processing and storage, but for those application aspects that need the lowest latency and most direct connection to the user – some processing can happen closer. “Closer” may be in a small regional data center, an individual enterprises’’ own facility, or on some cell tower somewhere. Wherever it happens, it’s about enabling that processing with the smallest footprint possible.
Arise MCP Edge
Given Mirantis’ experience in building compute platforms, and the ongoing trend towards doing some compute at the edge, it should come as no surprise to hear that the company is announcing an edge play. The Mirantis Cloud Platform Edge (imaginative name, right?) is a low footprint, low latency infrastructure play designed specifically for edge deployments. The product combines Kubernetes, OpenStack and Mirantis’ own infrastructure management solution, DriveTrain, to build a lightweight solution that covers the myriad of different operating paradigms (container, VM and bare metal).
But whereas some edge offerings exist merely as reference architecture or whitepaper hypothesis, the demo version of MCP Edge is available to download and use. What this means is that service providers can use the demo virtual appliance to deploy a Kubernetes-based, 6-node edge POP capable of running containers and VMs. An MCP Edge demo can be deployed on any server (or even a laptop) with a minimum of 24GB of memory. In addition, the running of mixed workloads is possible through the use of Virtlet, an open source container runtime interface pioneered by Mirantis and developed by the Kubernetes community.
Mirantis tells me that MCP Edge has been developed in partnership with its existing telco customers – names such as AT&T, Telstra, Vodafone, Reliance, STC, Cox Communications, and NTT.
MyPOV
I’ve said enough about Mirantis and its pivots, it’s time to look at the value of what Mirantis is offering here. There is no doubt that edge computing will be an increasingly important trend in the future. At the same time, deployment across a range of infrastructure approaches is also a growing trend. Put these two trends together and you certainly have a justifiable reason to suggest that MCP Edge is going to succeed.
It should be noted, however, that even the large public cloud vendors have been becoming more accepting of a continuum of solutions for different customers – even when this means that processing might occur somewhere other than their own big, beautiful and controlled data centers. All three of the big public clouds have been making forays into more distributed and “edge-ready” computing approaches.
As such, the question around MCP Edge is less one of whether or not edge computing will be needed in the future, but more one of whether a solution from a smaller vendor, as opposed to a packaged solution from one of the large cloud vendors, has a chance of success. It should be noted that AWS’ edge offering, Greengrass is downloadable and, in a departure from AWS’ usual approach, open source – what this means for MCP Edge will be an interesting journey to observe.
One needs to bear in mind that telcos, while owning a significant edge network already, also increasingly see themselves involved in more than simply offering dumb plumbing. A smart telco that sees the value in combining its existing network with some distributed processing, and offering that package up as a value-added solution to its own customers, would be doing much to secure its own future.
So I’m bullish about MCP Edge – it’s yet another way that open source cloud initiatives generally, and Mirantis in particular, are allowing large organizations to adopt approaches that bring the benefits of the cloud, without the disadvantages (as they see it) of being “all-in” with one particular public cloud vendor.
