Analyst Chat #85: Hybrid IT 4 - Hyperconverged, Edge and Cloud in a Box

Welcome to the KuppingerCole Analyst Chat. I'm your host. My name is Mathias Reinwarth, I'm lead advisor and senior analyst with KuppingerCole Analysts. My guest today is Mike Small, and we are finishing our series of episodes around the topic of hybrid IT and the many topics that are, um, in that, um, the headline "hybrid IT". We want to talk about future trends. Hi, Mike. Good to see you again.
Uh, hi Matthias, and thank you very much for keeping inviting me to these, uh, these wonderful things. So, um, in the past, we've talked about, uh, how the environment, the ways in which, uh, it services have been delivered has become fragmented with things in the cloud, things on premises and things around the edge, and that this has led to governance and management challenges. And the key approach for an organization is to find a common governance platform from which to, uh, govern the overall, uh, mixture of these things. So everything's going to get easier, isn't it? Well, except that, uh, the world never stands still. And so in this final episode, I'm going to talk about some of the things that are changing that may make things easier or more complex. And some of the things that, uh, the cloud service providers and others are trying to do in the hope of, in the belief that it may make it easier for you, but perhaps also following their own agendas.
Now, one of the, uh, interesting things is that, uh, a lot of the way in which it services are delivered is not actually from a data center that, um, uh, that there is now an awful lot of it. That's very important to businesses that sets in branch offices and, uh, in, in other places, uh, one very specific example of this is in fact, in the retail business, that if you go into any large, um, uh, supermarket, you can almost be certain that underneath the supermarket or in the back office somewhere, there is a very important piece of it. And it is what is keeping the tools running. And any of you, like I have been involved in this, see if you ever seen what happens when that it breaks on a Saturday morning, when everybody is trying to do that shopping, you will realize just the impact of norm availability.
And there was a famous riot in Hemel, Hempstead, where that actually happened. So now supermarkets are not the only place you have all kinds of different places that its really important and it must keep working. The, the modern factory is heavily dependent upon it to keep the production lines flowing, any kind of petrochemical plant depends upon it on premises. If you go out and drill for oil, although hopefully the world is not drilling for oil now, but if you go to do any of these things that are involved, uh, where you are extracting minerals, or you are trying to do things, uh, in, in, in the outdoors, you depend upon it. If you go to a foam, you will now find that, uh, in that foam, there is a dependence upon it. In fact, people don't realize just how much the, the, the modern day farming depends upon optimization in order to feed the world.
So what are you going to do about that? So this has led to, um, uh, th the, again, the, the problems of management and cell force of this. And so one of the interesting things that has happened is that the different environments in the different places lead to another dimension of complexity in this hybrid world that we've been talking about, that the world runs the application in the data center is a world that has to be managed, but usually slightly different to that bit of it that is running in the supermarket or in the factory. And it also needs to be managed. So, uh, how, how are you going to do this? Well, in fact, one of the interesting things that has happened is that the cloud vendors have started to move into this market, uh, and trying to persuade you to buy, uh, either a cloud in a box or the cloud software that you can install in the box. So all of the major cloud vendors are now providing either a, uh, a piece of equipment, which actually is managed like the cloud. And it exhibits all of the interfaces of the cloud that you can locate wherever you need to locate it, whether it is in your data center or whether it is in your factory or whether it is somewhere else that is important to you. And so they're even producing ruggedized versions of this that will survive in hostile environments like mine.
And so, uh, that is one of the, the, the key things. And the purpose of that is to try and provide a uniform environment from a point of view of management and, uh, governance. Uh, now, so interestingly, we have some vendors that, uh, some of the cloud services, like for example, Nike or Microsoft that have, uh, uh, the Azure stack. And with AWS, you have a range of things, including AWS outposts, both of which exhibit, that kind of characteristic, but allow you to effectively extend their cloud into your premises, um, and make various claims for what it can do. Now, on the other hand, you have, uh, another approach, uh, from, from the IBM cloud where they they've got IBM satellite, which is a piece of software, which you can install and they hope you could install in other people's other cloud vendors, clouds, so that they can give you that management that way.
So that is one of the approaches that is, is, is trying to, um, persuade people that that's a good way of doing it. And that allows you to solve some of the compliance problems that were causing problems, because if you have to have geographically located data, then you can ensure that is geographically located by actually physically controlling where that server is now at the same time, as that is happening, you have a mother completely different approach that is coming through the so-called hyper-converged infrastructure. So there is this sort of hyperconvergence, um, software that is, uh, uh, appearing nowadays like Terraform, which is sort of allowing you to build your cloud in your data center by, uh, having the software defined data center. And that is further being, um, uh, made complex because many of big hardware vendors have realized that actually in order to make it easier for people to buy their hardware without capital expenditure, you can now buy what is effectively, um, uh, a server as a service or storage as a service.
So what happens is you somehow are all given a physical piece of hardware that sits where you would want it to say, be it in an Equinix data center, be it in some kind of co-location data center or being in premises or in your, um, back office. And, uh, well, you buy a certain amount of capacity in this and what actually Dell or HP, or whoever delivers may actually exceed that. And, but you only get charged on a subscription for what you use. And moreover, what will happen is that the hardware vendor will say, well, we will give you a service level agreement, which says that if something goes wrong with that, well, first of all, we will monitor it. And if something does go wrong with it, then we will take responsibility for fixing it. So you don't have to, uh, raise the trouble tickets and chase and so forth.
That's all done for you now, has that made things simpler or has it made things more complicated, uh, in, in some ways it's changed the whole game. Um, and to, to sort of look at some of the things that you can say, if you are using a cloud service and you say, well, I've got this server over here and I've got some data in here. The cloud service invariably gives you a very simple API, which says, well, actually connect these things together or move them around and how they do that is cloud service, uh, intellectual property. If you try to do that yourself with any of these things, you find it it's actually really quite different, difficult even using these hyper-conversion for some. And you're now starting to get those tools being provided for, uh, the individuals to try and, uh, the individual organizations to try and do it.
So it, it, it looks good, but actually it's just even more complicated, uh, to, uh, to manage. Then if we go back to our story about, um, where, where earlier on, we were talking about how, uh, with our container based way of delivering, um, services that your, uh, components that deliver your service are all now located in different parts. So, uh, if you then look at the problems of management of the network in a container-based delivery of an application, you need a completely different level of network management to be able to optimize how, in fact, even to understand how the performance of that overall application is being determined by the network, the performance of a container-based application is no longer determined by the server. It will be determined by the network. And so there are now network management tools and network, um, uh, network devices that are being delivered by the big network vendors. So with that, uh, I'm going to say, thank you very much for listening to these, uh, to mind musings on, on the subject. And I hope that, uh, this has been helpful to your listeners.
Absolutely. Thank you very much, Mike, for, for sharing your experience and your insights also for just most recent research around the topic of how hybrid it evolves and actually has come full circle, uh, from on-prem to the cloud back to on-prem, but cloud on prem. Um, so this was really interesting to hear. Um, and I think this, um, evolution will continue and we will even see more developments there as well. And there are some around like, like fog computing and, and really putting workloads on, on, on smaller devices that are there any way cameras or whatever. So that might be a topic to look at, um, in the future as well. So not the controlled way of running, um, services on boxes and infrastructure provided by cloud service providers or hardware vendors, but really the more you get really our style work to do work, uh, way of doing things. So for that, thank you very much, Mike, for, for sharing that insight. And I'm really looking forward to having you soon in a further episode on that topic or a different topic. Um, it's always great to, to hear from your experience and from your research. Thanks again, Mike, actually, thank you. Bye-bye