The cloud computing era is still in its infancy but is much better understood than it was even a short while ago. As usual, though, people are asking “what comes next?” With infrastructure convergence, the next big thing is the rapid increase in distributed computing models as they pertain to the cloud.
There are a couple of important things worth considering when we think about that. One of the first things is in this whole area of convergence, where we are bringing all of these technologies together. We're creating an environment that needs to be highly orchestrated.
The importance of collaboration
There is definitely a need for increasing collaboration between the users of the network and the network itself. In the cloud world, there is currently a provider/subscriber model, but what we'll see going forward is a more collaborative model - a peering model.
The reason for this has to do with the massive amounts of compute power being embedded in everything from smartphones to enterprise switches. With that increase in computing power - in and out of the cloud - there is a need for an architecture for service delivery in a distributed model. Compared to the cloud computing model of today, the future distributed model will need to enable more collaboration between endpoints than the existing provider/subscriber model offers.
As we're already seeing in the market, there will continue to be more sharing of information about location and identities, and computing power will be used more for analytics and to create customised services. That requires a lot of information being exchanged, which in turn requires an established trust between the user, the network and the services. Privacy, integrity and encryption security needs are going to increase to establish the high level of trust necessary to make it work.
Moving to the distributed computing model
A lot of people look at the cloud today as reinventing the mainframe with more standards-based objects. The next phase is distributed computing, and with all this orchestration and trust establishment we'll start to move back to this kind of distributed model and we'll have a lot more need for this collaboration.
I like to think of the cloud almost as a mainframe model, but in a distributed model.
Lorna Hardie is HP Networking country manager, at HP South Africa.
To some extent, you can see this occurring already with converged infrastructures - like data centres - as we converge storage, network and servers. We have a need for all those IT departments to collaborate together to facilitate it.
What this means for channel partners
Channel partners can expect this to accelerate their ability to deploy and provide new services and solutions to their customers.
I like to think of the cloud almost as a mainframe model, but in a distributed model. What it means is there's a much easier ability to create, deploy and offer services and platforms that deliver those services without requiring the whole centralised procurement management model that exists in these large Internet data centres. It gives the channel partners a great opportunity to deploy incremental solutions that will fit nicely into this architecture instead of having to become part of a very large data centre cloud roll-out.
Additionally, it will also mean smaller partners will be more capable of handling the challenges of competing with one-stop shops that host every kind of service. While smaller channel partners today may find it very difficult to compete, the change to a distributed computing model via the cloud will alleviate some of their headaches in the future.
Share