Integrated switching and transport comes to the metro

Service providers historically have focused on increasing the capacity and service capabilities of transport and access networks to meet their customers' voracious appetite for more bandwidth and new services. To keep capital expenses (CapEx) and operational expenses (OpEx) under control, service providers have encouraged equipment vendors to innovate with a focus on integration. Initially this innovation referred to the integration of multiple service types into access—next generation digital loop carrier (NGDLC)—and, more recently, into transport platforms such as the multi-services provisioning platform (MSPP). 

However, there has been relatively little innovation regarding integration within the central office (CO). As a result, service providers are forced to spend an increasingly large percentage of their capital and operational budgets purchasing and maintaining banks of DSX panels, repeaters and tie-lines that provide connectivity between equipment in their COs. This problem is especially acute in COs that support large volumes of leased line and data traffic. Vendors began to address this problem in the mid to late 1990s with hybrid transport platforms, but now a new type of product, the MSPP core aggregation system (MCAS),* promises to fully address this problem.

The Final Frontier
A typical network consists of multiple COs that support services interfaces to the customer via access rings and interconnections to other COs via inter-office facility (IOF) rings. The quantity of access rings supported from a single CO varies from ones (in a small CO) to hundreds (in a large CO). The quantity of IOF rings is smaller, varying from ones to tens. Each of these rings is terminated in the CO by a Sonet add/drop multiplexer (ADM) or a MSPP. As a result, COs contain banks of Sonet ADMs and MSPPs.

CO aggregation and switching functions are supported by multiple types of network elements (NEs), each of which has been optimized to perform a single function. These functions include Layer 1 transport (Sonet ADM and MSPP); Layer 1 switching (wideband digital cross-connects, or WBDCS; and Layer 2 switching (Ethernet switch). The Layer 1 NEs are interconnected with DS-3 tie lines (see Figure 1 below). Large COs require repeaters to regenerate DS-3 signals that travel long distances (such as cable runs between floors). DSX panels are required to provide for the administration of and test access for the DS-3 tie lines.

Figure 1 PMO CO

These DS-3 tie lines, repeaters and DSX panels are essential to the operation of the network, as they provide the connectivity that binds the various components of the network together. However, these elements also represent a large and growing component of the service provider’s CapEx and OpEx expenditures. They are also the source of many service failures that result from human error, lack of redundancy and reliability issues inherent to coaxial cables and metallic connections. As such, the CO has become the last frontier for cost reduction for service providers and equipment vendors alike. 

Hybrid Transport Platform

During the mid to late 1990s, equipment vendors introduced a new type of transport product that supported multiple low-speed (OC-12) rings, referred to as subtending rings from a single high-speed (OC-48) ring. These NEs provided connectivity between IOF and access rings via a small, integrated STS cross-connect fabric that varied in capacity from 192x192 to 768x768 STS-1s, depending on the bandwidth and quantity of interfaces supported by the NE. Due to its support for multiple functions, vendors referred to this new breed of NE as a ‘hybrid transport’ platform.

Hybrid transport platforms effectively integrated a single high-speed and multiple low-speed Sonet ADMs with a small broadband digital cross-connect (BBDCS) system. This integration provided the following benefits to the service provider:

• A reduction in the number of DS-3 tie trunks, DSX panels and repeaters that would have been required to provide connectivity between a single high-speed and multiple low-speed Sonet ADMs. Note that DS-3 tie-trunks and their associated hardware were still required to support connectivity to other transport and switching equipment in the CO. This reduction in DS-3 tie lines reduced the number of truck rolls required to provision new circuits, reprovision circuits as a result of churn or network optimizations, and test or troubleshoot faulty circuits. 
• A reduction in the quantity of Sonet ADMs and the DS-3 interface cards that would have been required to support intra-office connectivity. This reduction of equipment resulted in real estate and power savings.
• An improvement in the reliability of DS-3 services due to a reduction in the quantity of Sonet ADMs and DS-3 tie lines (points of failure), and the resulting increase in customer satisfaction.

These benefits resulted in significant CapEx and OpEx savings for service providers’ small COs. However, hybrid transport platforms did not support sufficient transport or switching capacity to support the connectivity needs of large COs, and thus did not provide significant CapEx and OpEx savings for these larger COs.

In the late 1990s, equipment vendors enhanced their hybrid transport platforms to support integrated, private line Ethernet over Sonet (EoS) transport service. This enhancement provided point-to-point Ethernet connectivity between customer locations, eliminating the need for customers to purchase, administer and maintain specialized protocol inter-working devices (Ethernet to DS-1/DS-3 private line and Ethernet to frame relay).

Additionally, some private line EoS implementations supported limited Ethernet services aggregation capabilities—multiplexing data from two or more Ethernet interfaces supported from a single line card into a single Sonet STS container. This feature enabled service providers to realize limited efficiencies for centralized VLAN switching applications including increased transmission resource and switch port utilization. However, this architecture required that all data traffic be backhauled to a centralized switching location, even if the data was headed for a local destination.

Thus, private line EoS transport service provided an attractive alternative to legacy transport services for the customer and offered new revenue opportunities to the service provider, but it did not significantly reduce the service provider’s OpEx costs.

MSPP Core Aggregation Systems (MCAS)

In response to the service provider’s need for CapEx and OpEx relief for all COs, a new breed of product has been introduced, the MSPP Core Aggregation System (MCAS).

The MCAS supports multiple, multi-services line cards (scalable from ones to hundreds), with each of these line cards configurable for TDM, Ethernet or Ethernet over Sonet (EoS) services, and as an IOF or access ring node. When these line cards are configured for TDM services, the MCAS provides non-blocking connectivity between IOF and access rings via a broadband cross-connection fabric that is scalable from hundreds to tens of thousands of STS-1s ( see Figure 2 below). When these line cards are configured for Ethernet or EoS services, the MCAS provides VLAN switching between these line cards (see Figure 3 below). As such, the MCAS provides for the integration of the Layer 1 transport and switching functionality, as well as the Layer 2 Ethernet VLAN switching functionality, within the CO.

Figure 2 CO with MCAS

Figure 3 Ethernet VLAN Switching with MCAS

The integration of ones to hundreds of MSPPs, a BBDCS and a Layer 2 VLAN switch into a single NE enables a dramatic decrease in the number of intra-office DS-3 tie trunks and their associated DS-3 interface cards, patch panels and repeaters. Intra-office Ethernet connectivity requirements are also significantly reduced. These reductions in intra-office connectivity cabling and equipment result in a precipitous drop in CapEx and OpEx—as much as 60% to 70%.

The MCAS provides the following additional benefits to the service provider beyond those provided by legacy, hybrid transport platforms.

• The MCAS effectively integrates ones to hundreds of MSPPs and a scalable BBDCS into a single NE. In essence, the MCAS can be viewed as a BBDCS core supporting multiple MSPP blades (an MSPP integrated into each line card). This reduction of equipment results in additional real estate and power savings.

• The MCAS eliminates the need for service providers to backhaul Ethernet traffic to a centralized Ethernet switch, resulting in transmission efficiencies.

The MCAS also simplifies CO and IOF engineering and planning. With its high scalability in regard to both ring fan-out and switching capacity, plus with a high degree of interoperability, the MCAS can support ring additions and bandwidth upgrades well into the foreseeable future. As such, service providers no longer need to plan for switching matrix upgrades or the addition of additional boxes every couple of years. Extra capacity can now be added by simply adding additional line cards into the existing chassis. Additionally, since a single NE can provide connectivity between all IOF rings terminating at the CO, the IOF network can now be viewed as a large mesh supporting any-to-any node connectivity, thereby simplifying circuit routing and provisioning.

To fully realize these CapEx and OpEx savings and other benefits, the MCAS must support interoperability with a variety of NE types, particularly Sonet ADMs and MSPPs that are deployed on access rings and/or at the customer premises. Service providers have deployed legacy Sonet ADMs that support the OSI Seven-layer Stack on the Sonet Data Communication Channel (DCC) as well as next generation Sonet ADMs and MSPPs that support IP over the DCC in their networks. As a result, the MCAS must support inter-working with OSI and IP DCC implementations, as well as gateway network element (GNE) functionality. The MCAS must also support EoS inter-working, including ITU-T X.86 LAPS and G.7041 GFP inter-working, to support the full gamut of transport and access products.

MCAS clearly is the natural evolution of legacy, hybrid transport platforms, supporting the integration of ones to hundreds of MSPPs, as well as BBDCS and Ethernet VLAN switching functionality. This degree of integration enables a dramatic decrease in intra-office DS-3 and Ethernet connectivity requirements within the CO, leading to a precipitous drop in CapEx and OpEx. And, since the MCAS supports interoperability with a wide variety of legacy and next generation transport products, it can be dropped into existing networks with little impact to current operations procedures.

---

Bill McDonald is Manager-Product Marketing for Mahi Networks, Petaluma, CA. He can be reached at bmcdonald@mahinetworks.com.

Visit Mahi Networks online.

* (Note: Mahi Networks typically defines MCAS as "Metro Core Aggregation System," but for the purposes of this article, we use "MSPP Core Aggregation System" to underscore the concept of a system in which each line card in a chassis functions as an MSPP.")

return to article