The great migration

Most of the traffic coursing through today's metropolitan networks rides on Sonet-based rings. As anyone can easily surmise from the explosive growth of the Internet and related services, these metro networks are being stretched by surging traffic and the shift from voice to data services.

Although voice traffic continues to rise - albeit at a much slower pace than data traffic - voice-centric time division multiplexed Sonet rings soon will break down under the strain of voice and data traffic. Carriers struggling to meet these new demands are looking to break Sonet-imposed barriers in terms of layers of equipment and operational concerns - including network management and service provisioning - by using wave division multiplexing (WDM) to provide additional capacity and transparent transport for a variety of optical data services.

The good news in this situation - and there is some - is the new revenue opportunity for carriers that provide native gigabit Ethernet, Fibre Channel and Enterprise Systems Connection, or ESCON, services. However, the successful delivery of these services is contingent upon a network that has the required optical infrastructure, integrated network management across multiple technology platforms and integrated network design and simulation.

Unfortunately, most metro networks today fall short of these requirements.

Therefore, existing carriers need effective migration strategies to maximize the value of their current investments. No one would willingly "write off" existing infrastructures. Even emerging carriers that start fresh with all-optical networks must take heed: They will need to hand off traffic at the access, metro or long-haul areas of the network. Likewise, existing and emerging carriers need to be cognizant of the challenges associated with the integration of multiple technology platforms.

The status quo

Present networks are built on voice-centric equipment that is not optimized for other types of traffic. In addition, integrated management of the emerging WDM-based optical layer and end-user services still proves elusive because of the use of multiple technology platforms and the lack of management tools integrated across management functions and network segments.

With multiple technology platforms, service delivery can become difficult to oversee in metro optical networks because each platform has its own management system for performance monitoring and provisioning. When traffic goes over multiple technologies, exchanging information between these technologies can be challenging.

Despite many network equipment vendors' claims of having "open systems," the available network management systems still are proprietary. This requires service providers to expend a lot of energy to identify, troubleshoot and monitor performance.

In addition, existing management tools are designed to support individual network management functions. Standalone tools help network operators with network design, while other tools assist operators with day-to-day management. In this approach, network design is decoupled from actual network operation, and the operator must provide all the intelligence to optimize the design for the required operational performance, accounting for current network constraints.

Furthermore, current management systems are designed to manage individual network segments and provide all available resource data to the operator. The operator provides the intelligence to interpret and act on network data and to patch together networkwide views from individual segment information.

For example, if the operator needs to provision a new circuit, he or she must manually check for spare capacity in the network equipment. If there is no spare capacity, the operator must rework the network design and manually determine any new equipment and configuration requirements. Given the complexity of the network, this is no small task.

For example, in many environments today, a service provider that aims to deploy gigabit Ethernet service for a customer needs to perform these steps:

- Determine which classes of equipment are involved

- Manually calculate available network capacity on each equipment class and correlate against service

- Manually calculate additional equipment requirements

- Wait for additional network capacity to be installed

- Manually provision each individual service element

- Manually monitor service quality from each service element and correlate to the service level

Performance monitoring in existing systems requires individual performance tracking across multiple equipment. Consider a present-day architecture in which IP packets are carried over a Sonet transport platform. Performance is measured at the Sonet equipment and the IP equipment. Performance measures for Sonet equipment focus on the quality of the optical transport, while performance measures for IP equipment focus on the quality of the data packets received. In this example, the root cause of a problem in the IP equipment's performance actually may lie in the Sonet equipment.

For these reasons, it is imperative that the optical networking equipment community begin designing systems that will allow carriers to create a more seamless environment that will manage network connectivity, provisioning and management across multiple platforms.

This initiative should take into account equipment that supports point-to-point, ring and mesh network topologies with a seamless migration path among them. In addition, most carriers need provisioning and management systems that will interoperate with legacy operating systems and network equipment.

A next gen model

The desired next generation metro network needs to support the required variety of applications and end users as well as end-to-end service delivery across multiple technology platforms to reduce provisioning time and operating and maintenance costs. The challenge in deploying next generation metro networks is not providing more capacity, but rather providing users with services when and where they are required. Unlike long-haul networks that merely interconnect city backbones, metro networks must support a variety of applications and end users.

The next generation metro network will simultaneously transport inter-office traffic between metro central offices, connect businesses to other campus locations across a WAN, aggregate and transport consumer traffic and serve ISPs and other telephony carriers.

All these customers expect reliable, cost-effective transport that protects against equipment failure and fiber breaks. In addition, each of these end users may need to rapidly scale traffic requirements, add temporary capacity to accommodate special events or adjust user bandwidth according to the time of day.

Users will be hard-pressed to deploy these networks without comprehensive management tools to buffer network operators from the details and complexity of operating a network. Therefore, next generation metro networks should feature an end-to-end management platform to ensure seamless operation across all network elements and to integrate all network management functions - network design, configuration and operations. All this must be designed for ease of use for operators and customers.

Next-generation network management systems will allow reporting of configuration data to external operations support systems for flow-through provisioning, network surveillance, inventory control and advanced network modeling.

Next generation metro management systems should permit A-to-Z wavelength provisioning, which lets the operator work off a complete "top-to bottom" map of available services and paths, and point-and-click interfaces, which enables the operator to select and activate end-to-end circuits across the network. This will facilitate operations, administration, maintenance and provisioning (OAM&P) such as service activation and deactivation.

Carriers will achieve much faster turnaround because they need only know the service origination and termination nodes and the service type - including the access type and data rate - and will be able to see a graphical view of the network and services to point and reserve resources automatically.

Standards-compliant CORBA interfaces facilitate the automatic reporting or on-demand retrieval of service alarms to the network management layer. This emerging structure rests on the TeleManagement Forum's 509 interface agreements to promote interoperability between service provider applications, third-party applications such as those from Telcordia, and equipment vendor applications such as those from Lucent Technologies and Cisco Systems. These systems also will permit automatic inventorying and resource use reporting of network equipment across the network.

The next generation network management system will communicate with all network elements to track spare capacity and to provide real-time performance monitoring and alarm management. These systems will integrate network design, configuration and operations.

For design, the platform should provide integrated design tools for planning and optimizing new networks and upgrading existing networks. For configuration, the platform should provide detailed configuration procedures for new installations or incremental system growth. Finally, for operations, the platform should provide day-to-day network management capabilities, including end-to-end provisioning, performance monitoring, fault monitoring and isolation and inventory management.

To ensure ease-of-use for operators, the system must scale as the various network elements and the number of network elements grow. Furthermore, embedded intelligence should be used to provide the operator with concise management information across the network.

Similarly, to ensure ease-of-use for customers, next generation network management systems should allow customer self-management and self-monitoring of customer-specific network resources. Customer network management could be achieved through security administration which features partitioned, Web-based access to network resources through user-specific security profiles.

Any way they see fit

Both traditional Sonet and non-Sonet based traffic such as gigabit Ethernet, Fibre Channel, and ESCON are growing exponentially in every part of the network - especially in the metro segment. Optical networking provides the capacity and flexibility to deliver these services to end users. The challenges service providers are finding are determining the migration path to accommodate the new demands and managing all the diverse equipment involved in the multiservice network.

The use of multiple vendors to support customer-specific requirements allows network operators to use whatever technologies they see fit to accommodate those requirements. Integrated network management will allow network operators to leverage legacy infrastructure and to reduce the manual processes of provisioning and maintenance. And the costs that service providers recover through the use of network operations and management systems ultimately will translate to increased revenues.