Migrating to an Ethernet-centric Infrastructure

Good question. Many providers of 2G/3G mobile wireless backhaul are interested in converging their TDM and packet traffic over a single converged Ethernet network. In addition, many providers of TDM private lines are looking at ways to support incremental sales of private lines over their new Ethernet networks so they don't have to continue investing in legacy SONET/SDH equipment. To make this a reality, a few key standards have emerged to transport TDM over packet or Ethernet, commonly referred to as TDMoP. The IETF has defined two standards--Structure Agnostic TDM over Packet (SAToP) and Circuit Emulation Services over Packet Switched Networks (CESoPSN). Both are critical to achieve TDMoP especially if both channelized and unchannelized services are required. Essentially, they take TDM traffic and encapsulate it into a either Dry Martini or MPLS pseudowires for transport over Ethernet. By using pseudowires, the provider can leverage the OAM and network resiliency mechanisms. Another emerging standard is the MEF's CESoE which provides TDMoP without pseudowires. This approach is less mature. In terms of synchronization, this does present a challenge for TDMoP once native TDM connectivity is completely eliminated from the site requiring the TDM service (i.e. no physical T1/E1 to derive timing and synchronization from). So again, a few standards are emerging to overcome this challenge. One is from the IEEE called 1588 and another is from the ITU called G.8261. G.8261 specifies the timing requirements of Ethernet networks at the boundaries of Ethernet and TDM networks. It outlines the minimum requirements for network element synchronization using two methods for clocking distribution: Synchronous Ethernet and packet based methods (i.e. 1588). Synchronous Ethernet, as defined by G.8261, uses the physical layer of Ethernet and can only distribute frequency and not time of the day. It is not affected by impairments introduced by the higher levels of the network. IEEE 1588 is independent of the physical Ethernet layer and can distribute time of the day and frequency. However, it can be affected by impairments of the network such as packet delay variation.

Operations Administration and Maintenance (OAM) is a group of network management functions that provide network fault indication, performance information, and data diagnosis functions.

Support for Ethernet OAM capabilities is important for carriers providing Ethernet services to end-users as it allows carriers to detect and troubleshoot network faults remotely and monitor network performance. Ethernet OAM enables carriers to have end-to-end network visibility so they can quickly identify service issues and efficiently correct network faults without costly truck rolls to the customer site. As a result, services are easier to provision and maintain and conformance to Service Level Agreements (SLAs), that define the guaranteed Quality of Service (QoS) levels, can be verified

Ethernet OAM standards are still maturing. So far four key specifications have emerged from standards bodies:

• IEEE 802.3ah Ethernet in the First Mile (EFM) includes mechanisms for link operation and health monitoring, fault isolation, and remote CPE management


• IEEE 802.1ag Connectivity Fault Management (CFM) specifies end-to-end fault detection for Ethernet transport services


• ITU-T Y.1731 augments 802.1ag with performance monitoring capabilities


• MEF 10 Ethernet Services Attributes provides a standard framework for describing Ethernet services, including performance characteristics pertinent to SLAs