Video: Cloud Mobility: The adoption of Containers and Kubernetes Q2

What place should Kubernetes have in your strategic IT roadmap for the next say 3 years and what impediments might slow down their adoption?







VIDEO TRANSCRIPT

And the second question I have is what place should kubernetes have in your strategic it roadmap for the next say three years and what impediments might slow down their adoption?

So we'll start with andrew on this one,

Great thanks John i think um like any evolution um it's important to probably go back to the business requirements of why you would deploy something like kubernetes um again the kubernetes is a is an orchestrator um it is not necessarily a thing that runs your containers there's various run times for that uh as rays mentioned docker or container d and others that are emerging um and so to draw a comparison you would be looking at um you look at something like um the vmware vcenter would be kind of your equivalent of kubernetes and then you have your hypervisor is actually things that are running the workloads and so where this might fit into you into your strategy as more and more applications are being deployed or coming from the application developers uh in a microservice or containerized way certainly kubernetes is important to be able to orchestrate that and run those workloads in that containerized and dockerized fashion whereas previously you might have run an entire vm you'll now just sort of run just the individual micro services but i think also where there is value being added by considering a containerized approach orchestrated by kubernetes is the portability between clouds so the ability to run kubernetes anywhere uh and so that gives you that level of agility it also creates uh capability for competitive tension between uh cloud vendors for example so that you don't necessarily have to to railroad your application or services into a particular cloud you can actually take advantage of all the clouds uh and then the next evolution that will come after that we're hearing about is edge computing well again kubernetes can run workloads out towards the edge so i think by having kubernetes as part of your portfolio and transitioning the service and application deployment towards a containerized model that's orchestrated by kubernetes it really provides a huge amount of flexibility as you as you progress um into a more cloud cloud first type world um there is some additional enhancements that come off the back of kubernetes that would that almost be kind of day three um future future looking statements that i think ray is going to touch on around serverless and how that relates and we'll touch on that a little bit later on in the session i believe

All right brilliant Ray,

What so the place that kubernetes might you know the role the communities might play I imagine in your strategic grow mac for the next you know half of the decade really depends on the position that you're at right now and the velocity with which you're migrating away in your product life cycles or according to um you know your customer demands um if you're if you're already fairly you know down that path towards consistent environments for your developers and for uh or potentially as andrew said you're doing a multi-cloud ecosystem then you already understand um some of the pitfalls and and some of the the benefits involved um and i guess the big question as to uh how uh pivotal uh kubernetes might be in the coming years is going to be what are the the tools that you have right now that solutions exist in the in the helm package manager um ecosystem for example the helm being the the kubernetes native um package management service uh which is now adopting i think version three stable and has an enormous marketplace of ready to deploy in five minutes you can run a helm deploy command and you've got your mysql server you've got your elastic search cluster um auto scaling and configured according to best practices so if these are the sorts of applications that you already have in your stack then um this is a you'll be able to deploy uh secure and scalable solutions for that in pretty much no time and then yeah i don't know what um the impediments towards that are other than i guess uh familiarity and and um priorities i don't know much more to add on that question

Fair enough too okay Matt over to you,

Yeah thanks jonathan um problems the guys have covered off i think a bunch of a bunch of good topics problems we i think when containers were first you know came to bear and they were spun out of you know kubernetes was fun out of google um they were stateless right and and very much that meant that you know from from a container's perspective you know a container might be waiting for a file to arrive when a file arrives container is spun up it does something with the file container disappears and so containers will often use that kind of lightweight stateless manner but as we've increasingly used containers for more important business critical applications we've hit a challenge with stateful applications which mean that the data that is created by the application within that container needs to live on beyond the life of the container and containers didn't really think about that when when they first started and so you have quite a lot of complexity around you know how do i allocate a volume of data attach it to a container how do i make it the right size the right attributes in the right place and dynamically doing that is a real challenge and when you start to get at scale with the applications it becomes a major problem so you know from a netapp perspective we've been focusing very much on solving that problem we're a data management company so we've bought a whole bunch of open source solutions to market so that doesn't matter which cloud you're running your kubernetes in we can solve some of those problems for you because i think six out of ten applications you know the helm charts that that ray was talking about you know you either download and run mongodb in a container suddenly what happens to that data when that container goes and how do i replicate that to another container in another cloud and how do i make sure that data is consistent because now it's a database and it's got all my customer information in it so solving those problems is super important so they're a challenge and make sure you've got an answer to that as you start to adopt a container architecture