Opt for a smooth transition to Provider Backbone Transport (PBT)--Part II

Maximizing service resiliency
PBT provides carriers many new capabilities for traffic engineering, increasing scalability, and improving network reliability. Creating dedicated point-to-point paths through the network, however, is only part of the solution. From a subscriber's point of view, the only thing that matters is application performance and availability. Providing the high availability necessary for critical business applications and residential triple-play services requires much more than just redundant components and connections. The network must be aware of the services that it is delivering and possess the intelligence and ability to preserve the integrity of these services in the event of network failure or outage.

This characteristic of the network is known as service resiliency and requires that the concepts of five 9's reliability and 50 ms failure restoration response be extended up from the physical layer to higher network layers in order to improve overall availability of network services. Service resiliency is enabled by employing failover mechanisms at the network, hardware, and application layers. Failover mechanisms at the network level, such as supporting primary and secondary point-to-point dedicated paths, assure that traffic will always be able to find a way through the network. Redundancy at the hardware level guards against module and component failures in the transmission equipment. However, with more and more network outages being caused by software issues, modularity and redundancy in the operating system is now a requirement for a carrier class network. A modular operating system allows application independence, where a problem with one application will not be able to affect the performance of another. By addressing resiliency at the hardware, software and network layers, service providers can maximize service resiliency and provide real value to their subscribers.

Extending service reach
Service provider subscriber surveys have repeatedly shown that behind reliable service delivery, one of the key customer concerns with adopting a new service is the ability to reach all of that customer's locations. Hand-offs between carriers in traditional SONET/SDH networks are typically routed through a digital cross connect system. As the industry moves to an Ethernet based model, a new approach is also required in order to facilitate native Ethernet peering and the extension of services across provider boundaries.

In response to this the Metro Ethernet Forum is working on an Ethernet Network to Network Interface (E-NNI) specification that will allow for more efficient peering of Ethernet services between providers. The E-NNI specification will likely require the inspection of both the customer VLAN tag and the service provider VLAN tag in order to determine proper packet prioritization. This new specification has two important benefits. The first is providing a standard for the mechanics of exchanging packets and specifying service parameters between carriers. The second is that as a result of simplifying the interconnect, more peering arrangements will be made and Ethernet services should grow significantly as service reach expands. While not directly related to the PBT specification work, it is critical to choose an Ethernet solution with E-NNI capabilities when deploying a network today. This will simplify a future transition to PBT by providing the tools required for effective Ethernet peering to extend service reach beyond the provider's territory.

Cost-effective alternative to MPLS
Service providers with primarily SONET/SDH-based aggregation and metro core networks have been hesitant to move over to new infrastructure options because of the lack of control over traffic engineering and service resiliency. While MPLS does provide some of these capabilities, they tend to be very complex and expensive to implement. Additionally, whenever another service or new equipment is added to the MPLS network, it may need to be tuned again to provide adequate performance. Recent standards efforts in the MPLS camp have attempted to provide a more transport friendly and simplified version of MPLS to meet the needs of this application. However, a lack of carrier interest may cause an early demise for this approach.

Meanwhile, strong carrier demand for PBT stems from its ability to provide an elegant solution to adding traffic engineering and carrier class resiliency capabilities to Ethernet while providing seamless interoperability with existing PBB and Q-in-Q networks. Because it is built upon Ethernet-based technologies and uses familiar constructs, PBT does not require specialized personnel to implement and manage. Instead, PBT meets all of the customer requirements while maintaining the significant cost benefits of Ethernet.

Page 2: next page
Page 1 | 2