Preface

Optical networks were initially designed and deployed to provide high-capacity transport for carrying telephone calls. Today they are used to transport not only voice traffic, but also data traffic, including Internet traffic. In fact, it is the Internet that has driven the rapid and unprecedented growth of optical networks in the past several years. New technologies, such as Dense Wavelength Division Multiplexing (DWDM) and intelligent optical switches, have been the other catalysts of growth. Although this rapid growth of optical networks has ushered in a new era in high-speed communication, it has also created a new problem. Managing a large and dynamic optical network is a difficult task, and the existing network management infrastructure is inadequately tooled to handle it.

Service providers manage their optical backbones using Element Management Systems (EMSs) provided by the equipment vendors and Network Management Systems (NMSs) that are often built in-house. Both EMSs and NMSs use proprietary technologies and often need tedious manual interventions. Due to the proprietary nature of the management systems, it is difficult to quickly and cost-effectively integrate new technologies into the existing infrastructure. The intrinsically manual nature of the management systems increases the cost and the turnaround time of provisioning new services. It also increases the risk of inadvertent mistakes. The growing complexity of optical networks is making these problems even worse.

In order to better understand the complexity of the problem we are dealing with, let us consider a representative continental optical network shown in Figure 1-4 in Chapter 1. As shown in the figure, the network consists of multiple subnetworks at different levels of the network hierarchy. A number of network elements playing different roles, potentially manufactured by different vendors, and managed by vendor proprietary EMSs and NMSs add to the complexity of the network. Now consider the routine task of provisioning a connection between two offices of an enterprise located in City A and City B. The provisioning process in today's networks involves a number of steps, many of which requiring manual intervention:

First, the path has to be planned. This requires information on available resources in different subnetworks from City A to City B. Today, this process is mostly manual. The work order is broken up into multiple sub-tasks and assigned to different planning teams. The end-to-end path is planned in a piece-wise manner from City A to City B.

The connection now has to be provisioned over the planned path. Different provisioning teams responsible for different parts of the network execute this process. This requires configuring the network elements and the associated network management systems in the connection path from City A to City B.

The next phase is testing. It is now the job of the testing team to verify that the end-to-end service is actually working and all the databases (inventory, billing, etc.) are properly updated so that the service can actually be turned on and handed over to the customer.

The provisioning process described above has been intentionally simplified for the purpose of presentation. It gives an idea, however, as to why the typical turnaround time for provisioning optical services is several weeks to months. It also explains why even after a precipitous drop in network equipment prices, the cost of running a network is still very high. Note that service provisioning is only one aspect of network control and management. The other components of network management are equally cumbersome and inefficient. In fact, 80 percent of the cost of running a network comes from the cost of managing it. The bottom line is that in order to harness the benefits of technological breakthroughs, such as DWDM and intelligent optical switches, there needs to be a better way of controlling and managing the networks. Specifically, a gradual migration from an antiquated, semimanual network management infrastructure to one that is more intelligent and automatic has to occur.

To be fair, it should be noted that the control infrastructure seen in today's optical networks is there for a reason. The last generation of optical networking equipment was kept “dumb” for technology reliability and for time to market reasons. Also, static voice transport did not require the “on-demand” service expected by dynamic data networks and applications. Today there is both the need and the means to develop an intelligent control plane for optical networks. It was this observation that spurred vendors to create optical control planes for their equipment. The standards development organizations, such as the Internet Engineering Task Force (IETF), the Optical Interworking Forum (OIF), and the International Telecommunication Union (ITU), were quick to follow. This confluence of interest and influence has led to development of control architectures, protocols, and standards at an unprecedented speed.

Now that we have made a case for the technical desirability of the optical control plane, does it make business sense? In today's difficult telecom environment are service providers willing to make the investment to deploy a dynamic control plane? While no one has a definitive answer to this question, a strong case can be made for the value of the optical control plane from the point of view of “return on investment.” Most of the cost of running a network today comes from operating expenses (opex). Due to the growing size and complexity of the network, this cost is actually increasing both in absolute and relative terms. In fact, many experts predict that even if service providers freeze their capital expenditure (capex), the ever-growing opex coupled with a decreasing revenue stream will seriously erode their profitability. The only way service providers can come out of this vicious cycle is by reducing their opex through better automation. Deployment of the optical control plane can help in this process. So, although it is an investment that service providers have to make, there seems to be a good reason to make that investment.

Notwithstanding these arguments, it is clear that the optical control plane is not going to be widely deployed overnight. Like any new technology, the adoption of the optical control plane will be gradual and in phases. This evolutionary process is probably a good idea both from technology and business standpoints. The control plane is still a fairly new and an evolving concept and hence gradual deployment will mitigate the risks associated with deploying a new technology. From the business standpoint, a phased deployment, with an objective to maximize the “bang for the buck” is a likely strategy.

In this regard, a number of service providers have started deploying intelligent optical networks and the associated optical control plane. Most of these early deployments, however, use vendor-proprietary protocols and software. Slowly but surely these proprietary technologies will evolve to become standards compliant. Evidence of this can be seen from the trials and prototype deployments of control plane standards developed by the IETF and the OIF. This process is likely to accelerate as service providers feel more pressure to upgrade their operations and management infrastructure to control the runaway operational cost.

The optical control plane has the potential to be the next most important development in optical networks. Consequently, it is an important topic with which service providers, vendors, and the business community dealing with transport network should be familiar. We have been intimately associated with optical control plane right from its genesis, not only as technologists but also as practitioners involved in the development and the deployment of this technology. In this book we share our firsthand experience in developing the architecture, protocols, products and standards for intelligent optical control plane. The book has been organized as follows. In the first few chapters, we review the existing optical networking technology, architecture, and standards. Our focus is on Synchronous Optical Network (SONET) and Synchronous Digital Hierarchy (SDH), which are the predominant technologies in this area. We then introduce the modern optical control plane and describe its components in detail. Next, we address the interaction between the optical control plane and the existing network management systems. We conclude with a discussion of control plane interworking, including the state of the standards and deployment.