I have just spent a week in Korea and Japan meeting with customers on the subject of Cloud. I can tell you, the subject is red hot out there right now. But what is really interesting is that it got me thinking about what we offer and how we typically focus it on the parts we know well. Indeed, meeting with CIO’s we talk about the resources (servers, storage, networking, services, applications) we require delivering cloud services, but there are a number of areas we overlook.
This became clear to me when we started discussing the need for R&D clouds. Many companies are looking at this space to boost the productivity of their engineers, allowing products to be brought to market faster. Now, who speaks about R&D starts talking about the use of graphical images and the need to transform, rotate, zoom and change them interactively. And then we all complain about bandwidth, about how slow the networks are and how this stops us of using the cloud for such activities. Well, is that completely true?
Technologies exist today to reduce the latency of networks, to ensure high throughput etc. By combining staging servers close to the end-users with advanced networking capabilities we create an environment where such services can be delivered. Why don’t we address that, rather than just complaining about it?
It’s probably because most companies do not run their networks and as such consider them outside their scope of control. This may well be true, but is that a reason for not addressing the issue.
I believe telco’s have a unique opportunity today to work with enterprises in these areas. Not only will it help them increase the traffic in their networks, but it will also allow them to foster long-term relationships with their customers.
When interviewed about their fears concerning the public cloud, CIO’s are typically highlighting security and availability as two key issues. Traditional public cloud providers do not provide satisfactory service level agreements. Telco’s wanting to deliver cloud services to their customers can easily do that by using high speed interconnects between their datacenters. They can run that on their existing fiber channels with limited extra investments, making 99.99% SLA’s in reach of any customer. Also as they are local, they can ensure those datacenters are within the appropriate boundaries for compliance, which is something the traditional public cloud providers are not able to do.
But that is not all, when we consider special clouds, such as R&D clouds, owning the network infrastructure is ideal to deliver solutions to the problems described above. Replication servers can be located where they have their switches, limiting the extra investments required. High bandwidth connections often already exists between switches and spare capacity can be used to synchronize the R&D files with the cloud environment. Fiber to the home, which is the next step in our evolution in network connectivity, will facilitate the delivery of fast images to the home.
Alternatively, technologies such as remote graphics, where the graphics workstation remains in the datacenter and pixels are transferred over the network lower the bandwidth requirements. However, such approaches involve the inclusion of special resources (in this case blade workstations) in the cloud environment itself.
Today, digital TV is already provided to many homes around the world. It’s a good start. To take this further, cloud service providers, telco’s, equipment providers and end-users will have to work more closely together to take full advantage of the technologies already existing and move them further.
The unique position of the telco’s, owning the networks, places them at the center of that ecosystem. Unfortunately, many of them do not seem to grab the opportunity as they do not really know how to evolve their cloud offering. The ones that do are the winners of tomorrow.
