14.5. Deployment Issues
The recent turmoil in the telecommunication industry has clearly altered the scope and the time frame of deployment of the optical control plane. In the early part of 2000, some predictions called for widespread deployment of intelligent optical networks containing hundreds or thousands or nodes, each with thousands of ports. This type of deployment, however, has not occurred so far. In fact, recent deployments of intelligent optical networks have consisted of smaller networks with perhaps tens of nodes. At the same time, some of the more sophisticated new applications that had been expected to be widely deployed are slowly entering the trial and testing phase. Several important issues listed below need to be resolved before wide scale deployment occurs.
Interoperability with legacy infrastructure: Today's transport infrastructure consists of a huge base of legacy SONET/SDH based network elements. Any deployment of optical control plane has to coexist with this legacy infrastructure and gradually evolve with it. It requires careful planning, extensive testing, and a well-orchestrated evolution strategy to make sure that this transition is smooth and does not affect service.
Interoperability with management infrastructure: Current transport networks are supported by legacy operations support systems (OSS) consisting of multiple layers of NMSs and EMSs. Any deployment of the optical control plane will require integration with the existing OSS and management systems.
Maturity of standards: Standardization of the optical control plane is still far from over. Some components, such as provisioning, are more mature and have received closer attention than others, such as restoration. Certain other components, such as interdomain routing are still in early stages of development. The proposed standards must be tested more extensively and must undergo broad interoperability trials before final adoption and deployment.
Operational experience: Most of the protocols that constitute the optical control plane have their roots in IP networks. The operational environment and requirements of transport networks are very different from those of IP networks. Transport network operators might want to develop some experience in using these protocols in limited settings before deploying them widely.
As a result of these conditions and the prevailing economic environment, the deployment of optical control plane is likely to be more evolutionary than revolutionary. Specifically, operators are likely to deploy components of the optical control plane that give them the biggest “bang for the buck.” For instance, automatic neighbor discovery may first be integrated with the existing provisioning and inventory maintenance systems. Similarly, interoperable signaling for restoration (and perhaps provisioning) may be deployed before distributed routing is implemented.
In any case, deployment of the dynamic control plane in some shape or form seems almost inevitable. Its long term benefits outweigh the short term challenges. The architectural model based on control domains eases the gradual introduction of the optical control plane in carrier networks. As the standards mature, the operators are likely to get more exposure to this new technology and appreciate its benefits. Also, with time, data traffic will further increase, and the business case for a dynamic optical control plane will become stronger. So it is not a question of whether—but when—the optical control plane will be widely deployed.