Archives

New Architectures, New Capabilities

By Esteban E. Sandino, AT&T Broadband

Supporting more traffic in the metropolitan broadband network is easier with approaches that migrate from dense wavelength division multiplexing (DWDM) architectures to emerging transport technologies.

The delivery of advanced interactive services continues to pose new challenges for broadband operators. Hybrid fiber/coax (HFC) access and last-mile distribution networks must support the increased bandwidth residential customers demand, so operators have deployed deep-fiber HFC architectures to meet these needs.

Operators also require significant bandwidth for metropolitan broadband transport networks that manage and distribute content associated with high-speed data (HSD), HFC telephony, video-on-demand (VOD), interactive television (ITV) and other services feeding the access networks. Advanced planning and provisioning to support increased traffic levels is as critical in the metropolitan broadband network as it is in the access and distribution networks. Operators need to explore new architectures and technologies that optimize their fiber resources.

Metro challenges

Early metropolitan broadband transport networks were deployed primarily for distribution of broadcast video. Installed digital transport systems were proprietary and offered only limited support for voice and data services. Synchronous optical network (SONET) rings often were used in parallel to transport delay-sensitive voice services. Moreover, additional processing layers--in the form of asynchronous transfer mode (ATM) switches or other Layer 3 hardware--also were deployed to transport and deliver packet-based data services.

One obvious disadvantage to this approach is that available fiber and bandwidth are not optimized across platforms and services. Another major disadvantage is the need to operate and maintain multiple signal processing and transport platforms. More often than not, supporting emerging packet-based services over these architectures requires separate provisioning of transport bandwidth in the form of point-to-point circuits. This process is followed by separate Layer 3 hardware configuration to ensure associated traffic bursts and quality of service (QoS) are handled properly. As the share of packet-based traffic over these networks continues to expand, the network's limited scalability is a growing concern.

However, new transport technology options have emerged and triggered a fresh approach to metropolitan network architectures. These new options are in the areas of DWDM and multiple service transport platforms.

DWDM in the metro network

DWDM technology already has been deployed in HFC networks to support segmentation of service areas, but its deployment has been minimal to non-existent in the metropolitan network. This situation is changing as support for operation in the 1550 nanometer (nm) International Telecommunication Union (ITU) wavelength grid continues to spread among transport platform vendors, and the cost of ITU laser sources and DWDM components continues to decrease.

Using optical DWDM technologies, an operator may activate multiple ITU wavelengths per fiber to immediately address fiber bottlenecks. One approach to DWDM in the metro network is to initially activate a minimum number of ITU wavelengths per fiber at 200 GHz (1.6 nm) spacing.

This approach supports immediate consolidation of multiple legacy transport platforms (see Figure 1). DWDM coupler technologies available today are able to support initial ITU wavelength activation at the 200 GHz spacing, while providing a migration path to 100 GHz spacing without service interruption. Moreover, the eventual introduction of tunable laser technology will reduce ITU laser sparing levels and inventory costs, thus facilitating migration to 100 GHz and even 50 GHz spacing in the metro network.

Advanced metro platforms

Advanced metro transport platforms will support not only legacy applications--such as digital telephony, analog/digital broadcast video and SONET-framed Internet protocol (IP) and ATM services--but also the formats and protocols of emerging packet-based interactive applications. A minimal set of supported features will include:

• Low delay and jitter;
• Time division multiplexing (TDM) interfaces, such as Digital Signal, level 1 (DS-1) and DS-3 for telephony service transport;
• SONET-type interfaces, such as Optical Carrier 3c (OC-3c), OC-12c, and OC-48c for legacy data service transport;
• High-speed optical aggregation interfaces: OC-48, OC-192 and higher;
• Moving Picture Experts Group 2 (MPEG-2) transport stream distribution via Gigabit Ethernet and 270 megabits per second (Mbps) Digital Video Broadcast-asynchronous serial interface (DVB-ASI);
• Native Ethernet interfaces, such as 10 BaseT/100BaseT/Gigabit Ethernet;
• Layer 2 protocols, for example, 802.1p/Q for class-of-services and virtual local area network (VLAN) support, media access control (MAC) learning, 802.1d spanning tree protocol and so on;
• Layer 3 protocols, such as border gateway protocol (BGP), open shortest path first (OSPF), routing information protocol (RIP) and so on;
• QoS-supporting mechanisms, such as traffic classification, tagging, queuing, forwarding, rate shaping and congestion management; and
• Dynamic bandwidth management and prioritization mechanisms.

In addition to the listed physical interfaces and protocols, SONET-type traffic protection and recovery mechanisms continue to be a critical feature in the new transport platforms. This status is reflected in the requirement for sub-50 millisecond (ms) automatic protection switching (APS) to support traffic restoration in case of failure in fiber ring architectures.

Various vendors are addressing metro broadband network requirements. Emerging platforms generally fall into two broad categories. The first are so-called multiservice provisioning platforms (MSPPs) based on current SONET standards. The second are pure packet-based platforms implementing resilient packet ring (RPR) architectures and protocols, or variations of this technology.

SONET MSPP platforms

SONET MSPPs incorporate high-density TDM switching matrices and separate packet processing engines implementing Layer 2 and Layer 3 protocols, for processing and routing of IP and other packet-based traffic. Transport bandwidth is allocated among TDM and packet-based services using granularities as small as 1 Mbps for more efficient bandwidth usage. This allocation presents a significant advantage over traditional SONET fixed pipes.

SONET MSPPs fall into two sub-categories. The first category supports full TDM traffic protection features, plus added statistical multiplexing capabilities for packet-based data traffic. Time- and delay-sensitive voice and video traffic is transported over standard TDM channels. Packet-based traffic is encapsulated, framed and then statistically multiplexed into fixed-size clear OC-3c/OC-12c/OC-48c channels.

The second category of MSPP implements a streamlined version of SONET that supports basic framing and functions, such as failure notification. In these platforms, all traffic--TDM and packet-based--is encapsulated in frames prior to transport over an optical channel at SONET rates.

Packet transport platforms

Packet transport platforms mostly implement RPR or some variation of it. A generic RPR architecture consists of two counter-rotating optical fiber rings interconnecting a number of RPR nodes, with both rings carrying working traffic. Ring bandwidth is shared across all nodes and its allocation is optimized for high-burst, variable-rate packet traffic. Congestion- avoidance mechanisms ensure near full usage of available bandwidth on both rings.

RPR features include allocation of service bandwidth in small increments; stripping of unicast packets once a packet reaches the destination node to support bandwidth reuse in subsequent ring segments; drop-and-continue to support service broadcast and multicast; and packet time-to-live and class of service (CoS) indicators. Some implementations also support legacy voice and other delay-sensitive traffic via TDM circuit emulation. Layer 2 and Layer 3 protocols also are incorporated in these platforms.

A major feature of RPR architectures is that, unlike SONET, each RPR node is topology-aware, and knows of more than one path to reach another node. This design supports logical meshed networks as opposed to point-to-point nailed-up connections. Protection mechanisms based on traffic re-routing at the source nodes enable service restoration in sub-50 ms timeframes, while preserving all service connections for high-priority traffic. RPR technology is under standardization within the Institute of Electrical and Electronics Engineers (IEEE) 802.17 working group.

Deployment strategy

How these technologies are deployed will depend on individual network needs. A potential strategy would involve initial deployment of DWDM technology to achieve immediate service consolidation at the physical fiber layer. Although this strategy will not address bandwidth optimization across services, the basic DWDM infrastructure may be the first step in the migration of all services to advanced metro transport platforms.

Advanced transport systems may first operate over their own set of ITU wavelengths and coexist with legacy systems. As more services migrate to the new platforms, an operator will start taking advantage of simplified service provisioning and management. ITU wavelengths may be released and reallocated to other uses as services are consolidated.

The integration of DWDM and advanced transport technologies will support service consolidation. It also will position the metro networks to benefit from further advances in optical switching technology and protocols, such as generalized multi- protocol label switching (MPLS) that promise to eventually support a common control plane for TDM, packet, wavelength and fiber services.

Concerted approach

To support advanced services, operators need a concerted approach that addresses not only the bandwidth needs of the access networks, but also the higher demand for bandwidth around the metropolitan networks that feed the distribution layers. One strategy is to first establish a common DWDM infrastructure to support both legacy and advanced metro transport platforms. This plan also should be coupled with the deployment of common management and provisioning platforms to simplify network operations and support service scalability with minimal network disruption.

Esteban E. Sandino is executive director, networking, for AT&T Broadband. He may be reached via e-mail at .

 Back to September 2001 Issue

Access Intelligence's CABLE GROUP
Communications Technology | CableFAX Daily | CableFAX's CableWORLD | CT's Pipeline
CableFAX Magazine | CableFAX databriefs | Broadband Leaders Retreat | CableFAX Leaders Retreat