Carrier SDN has been a hot topic lately. It seems that every transport and routing vendor is now on-board with the idea of separating the control plane from the data plane, as well as centralizing the control plane.
Another important criterion defining Carrier SDN is having open and standard interfaces on network equipment elements (i.e. WDM and Router systems) to ease the communication between it and the SDN Controller. This is where vendors are opting between three choices.
The three choices that vendors have are as follows:
- First choice: enhance their products with the use of open protocols and interfaces such as OpenFlow.
- Second choice: expand their network or element management system (NMS or EMS) to enable some of the qualities that an SDN Controller has, as well as include open APIs to permit communication between it and third party SDN Controllers.
- Third choice: offer an entirely separate product (software running on a standard x86 server) that performs a number of additional tasks, as well as act as a bridge between the SDN Controller and the vendor’s network equipment and NMS.
No matter which route the manufacturer chooses to permit centralized control of their equipment, the service provider will obtain the building blocks to plan and roll out a new network architecture based on Carrier SDN. And why wouldn’t they?
Carrier SDN is an architecture that will enable operational improvements for service providers resulting in lower operational expenses and faster time-to-market. This is obtained because Carrier SDN will enable service providers to architect a network that will centralize the control plane of all the OSI network layers (L0 through L3) and generate a real-time abstract view of the network. With this a service provider can perform traffic engineering and path computation faster with fewer resources. Furthermore, with the higher integration among layers, the amount of network capacity set aside for protection can be better managed, increasing network utilization. In addition to lowering operational time and expenses, we believe Carrier SDN will reduce service providers’ time-to-market of new money-making services, increasing their revenues and maximizing their profits. So when will we see the first Carrier SDN roll-out?
Well, we’re at the beginning. There haven’t been any large global scale roll outs, but there have been some meaningful deployments such as China Telecom with Huawei and PacNet with Infinera. Both are starting out at the transport layer, but we anticipating future deployments will include the packet layer.
I can’t wait…