Some service providers are content to keep up with the Joneses, launching Infrastructure as a Service (IaaS) offerings to show customers they have a cloud play. But for providers focusing on more advanced, higher-layer services, today's concept of cloud computing is already outdated.
In this video blog, cloud computing and networking expert Tom Nolle outlines three goals providers need to accomplish if the cloud is to evolve. Be sure to also read the companion article to this video, New vision of cloud calls for new cloud architecture design.
The following is a transcript of the video blog.
Hi, I'm Tom Nolle, President of the CIMI Corporation. Everybody is interested in the cloud, and if you're a potential cloud provider, you're especially interested, because it's what you hope to make money on. And a lot of people think that the cloud opportunity is passing them by, but the good news is that only about 1/250 of cloud opportunity has really been realized. The bad news is that we're not going to realize it by thinking about the cloud the way that we're thinking about it today.
We really need to re-conceptualize the cloud and re-conceptualize our vision of the cloud market if we want to make it successful. And there's really three elements that have to be focused on here. First of all, the cloud has to be a real IT architecture, and not just a mechanism for remote hosting. We can do remote hosting now. We can do virtualization now. If all we're going to do is server consolidation to remote hosting, there's no real opportunity here to be addressed. What we really need for the cloud is a Platform as a Service concept that allows us to write applications that are designed to be cloud-distributed. And if we can do that, then those applications can add significantly more value than applications that were never designed to be used on the cloud and are somehow kind of shoehorned into the process.
The second thing that we need is software-defined networking. You may not have heard of SDN's as being connected with the cloud, but software-defined networking is a model that provides a mechanism for software control of network behavior. And so, if we're going to write software that's designed to be cloud-distributed, it follows that that software needs to be network aware and needs to be able to control the way that network services are consumed by it and for it, by users. And SDN is the logical way to encapsulate this notion of cloud control of networks into 1 workable concept. It still needs work, but it's there and it's an available tool for us.
Then the final thing about the cloud is plug-and-play, or as the service providers would call it, "federation." You can't tack federation on any more than you can tack security on. An application architecture has to be secure and it has to be plug-and-play, meaning it has to support resource independence and application migration and mobility. And we have to build that in to the cloud. And in order to do that, we have to think about federation as an element of the cloud platform and not as an adjunct to it, not as a separate set of tools or a separate set of interfaces, as we conceptualize them now.
If you look at all of these things, we're not very far, in truth, in any of these areas. We do have Platform as a Service players from people like IBM and Microsoft and, to an extent, HP. OpenStack is developing into at least a potential for a PaaS. We do have SDN evolution today, but we don't really have strong integration with it in the cloud, and we've done very little with federation. We hope to be able to keep you up-to-date with the developments in these areas, and we hope that you'll stay tuned and continue your interest in this topic. Thank you. I'm Tom Nolle.