The explosive growth of IP in residential triple-play, mobile/wireless, and business data applications has triggered a global networking migration to Carrier Ethernet transport infrastructures. As part of this movement, network operators find they must expand the capabilities of their current transport networks to accommodate high-growth and bandwidth-intensive Ethernet-based services.
Initially, carriers tried to meet this need by implementing native Ethernet switching over fiber, but soon determined that they could not achieve the required scalability; operations, administration, and maintenance (OAM); sub-50-msec resiliency; or support for legacy services such as TDM and ATM. Then, in an effort to become more “carrier class,” Ethernet transport became largely based on Ethernet over SONET/SDH. This enabled service providers to deploy Ethernet reliably and quickly but failed to meet increasing bandwidth demands and the need for greater service flexibility and improved economics.
As a result, a new approach to Carrier Ethernet transport has emerged that enables these networks to evolve beyond native and SONET/SDH-based architectures found in first- and second-generation deployments. This new Carrier Ethernet transport network is supported by three key technology pillars — Optical Transport Network (OTN)/G.709, pseudowires, and intelligent optical networking — to address the future requirements of Ethernet-based networks and services.
Carrier Ethernet transport has emerged as an effective way to address the bandwidth and economic requirements of new corporate data, residential triple-play, and 3G wireless data services. Service providers are confident that these new services will help them grow revenues and offset declines in traditional products such as voice and private lines. Analysts predict that business Ethernet, mobile data, and IPTV services will represent an opportunity of more than $50 billion for carriers in North America alone.
Traditional Carrier Ethernet transport methods, however, do not effectively address the anticipated growth or requirements of these new services. Ethernet over SONET/SDH networks that provide dedicated, point-to-point connectivity for Ethernet traffic are unable to meet the need for multipoint connectivity to accommodate the any-to-any nature of IP. Through the use of mesh topologies and pseudowires, Ethernet transport networks become well suited for multipoint-to-multipoint services.
In addition, the tremendous increase in bandwidth requirements found in triple-play networks often requires up to 25 Mbits/sec per subscriber for IP video alone. The ability to manage and scale such bandwidth requires a new, forward-looking architecture leveraging optical technologies like OTN and WDM. These technologies not only scale better than alternatives but offer carrier-class management — a critical element to lowering operational expenses.
Ethernet was originally designed and built as a networking technology for use inside the LAN, where the number of users and end-to-end management could be tightly controlled. Once it escaped into the WAN, that control was lost and many features that make Ethernet desirable became marginalized. In order for Ethernet to become truly carrier-grade, transport equipment must overcome these hurdles to address the needs of service provider markets.
• Scalability: Ethernet needs to scale to handle high volumes of traffic that have long, continuously high utilization rates. On a given interface, Ethernet must be able to scale to tens of thousands of virtual LANs (VLANs) and be configurable across the range of 1 Mbit/sec to 10 Gbits/sec and to 100 Gbits/sec in the future.
• Quality: Ethernet must support committed bandwidth and burst thresholds while minimizing latency. It must be offered at the same level as traditional variable bit-rate services such as data virtual private networks (VPNs) and LAN extension. Ethernet must support multiple traffic contracts and quality-of-service (QoS) classes to match the characteristics and requirements of the services it is transporting.
• High reliability: Ethernet must be implemented with high availability and no single point of failure through both carrier-grade hardware and software. It must also meet or beat SONET/SDH recovery times of 50 msec with transit latencies in the tens of microseconds per switch-hop.
• Service management: Ethernet must be simple to manage and leverage existing tools, management systems, and OAM methods. It also needs to easily integrate into existing back-office operations.
• Compatibility with existing technologies: Ethernet must interwork seamlessly with legacy networks and services and be able to transport circuit-based traffic over packet infrastructures.
Since traditional products available in the Ethernet market were designed either for best-effort LAN traffic or simple bit transport, they lacked the ability to handle the combined challenges of scale, quality, reliability, service management, and TDM-to-packet conversion that must be met for Ethernet to be a true carrier transport mechanism.
To overcome the deficiencies of current-generation Ethernet transport networks, leading standards bodies developed enabling technologies to help build the foundation of a new Carrier Ethernet transport network. These include pseudowires, specified by the Internet Engineering Task Force (IETF), which emulate Layer 1 and Layer 2 services over IP/MPLS; ITU G.709 OTN for multiservice transport and SONET/SDH-like management with integrated WDM management to scale to 40 Gbits/sec and beyond; and intelligent optical networking standards G.ASON and GMPLS, for automated service provisioning and self-healing networks.
As shown in Figure 1, this new Ethernet transport architecture overcomes the challenge of supporting high-growth residential triple-play, business data service, and mobile wireless backhaul applications. It provides cost-effective metro transport with Ethernet aggregation intelligence and backhauling of traffic over the OTN infrastructure to a centralized service edge for delivery to the MPLS or SONET/SDH core.
Let’s look at these pieces individually. The key to understanding OTN is summed up in one word: transparency. For this reason, OTN (a.k.a. “digital wrapper”) is the ideal standard on which to support not just Ethernet, but many protocols over a converged network.
OTN specifications provide robust management analogous to SONET/SDH. But unlike SONET/SDH, OTN can carry a much larger payload and guarantee delivery of the underlying payload with its own performance management data intact. For example, OTN can fully encapsulate an OC-192/STM-64 or even a 10.7-Gbit/sec Ethernet LAN PHY frame and guarantee delivery across multiple OTN networks. OTN transparency does not end with SONET/SDH and Ethernet. The same benefits also apply to asynchronous data services like Fibre Channel, ESCON, and FICON that lack physical-layer performance monitoring capabilities and fault isolation necessary for a high QoS (see Figure 2).
The inherent flexibility of OTN is enabled by extending transparency to the timing plane. This allows the mixing of both synchronous and asynchronous signal types on a common wavelength. Moreover, synchronous services with different clock sources can be transported side-by-side, something not possible with SONET/SDH.
Another method of improving the cost structure of Ethernet is to reuse existing carrier infrastructure. Pseudowires enable reuse of a service provider’s significant investments in IP/MPLS to handle the growth of Ethernet. Through the use of pseudowires they can efficiently transport Ethernet over MPLS because pseudowires, not unlike OTN, provide transparency of Layer 1 and Layer 2 traffic over a Layer 3 IP network.
Pseudowire technology, which the IETF standardized via the Pseudowire Emulation Edge-to-Edge (PWE3) sub-working group, originated as “draft Martini” and is now defined by RFC 3916 and other ongoing drafts. As the name implies, pseudowires emulate the essential attributes of Layer 1 and Layer 2 services across a converged packet-switched network, most commonly IP/MPLS, as shown in Figure 3. Through the use of pseudowires, Ethernet and other non-IP traffic is sent across an IP/MPLS network while ensuring the service characteristics of that traffic remain intact and uninterrupted.
Because Ethernet is the new transport technology of choice it must support non-Ethernet traffic to truly achieve widespread deployment. Pseudowires enable a service provider to transport legacy TDM, Frame Relay, and ATM traffic across the Ethernet transport network.
In addition, pseudowires can be extended to the access network using a method called “Dry Martini.” Dry Martini encapsulates traffic into a pseudowire and transports it across the metro OTN infrastructure for delivery to the service edge, where the pseudowire is terminated or switched across an MPLS core.
Pseudowire technology delivers many benefits to the new Carrier Ethernet transport network including OAM, high reliability with strict QoS and network resiliency, scalability, multiservice support with Ethernet interworking, and industry-proven interoperability.
Finally, an intelligent optical network is a key enabler to improving reliability and lowering the cost of Carrier Ethernet transport. The benefits of intelligent optical networks are easily quantifiable with analyses of capital and operational expenses that reveal significant savings when compared with traditional SONET/SDH networks.
Intelligent optical networks increase efficiency in resource utilization, automate service provisioning and inventory update processes, and deliver high service availability. The intelligent optical network is enabled by control-plane functionality defined by the ITU G.ASON standard, which relies on a suite of signaling and routing protocols that support autodiscovery of network elements (NEs) and resources, the setup and teardown of light paths across the optical network, and mesh restoration.
Looking at the intelligent network concept in more detail, a key factor in lowering the cost of Ethernet transport is end-to-end automated services provisioning. The intelligent optical network rapidly responds to service requests by provisioning network resources automatically, across single- or multitechnology networks. Ethernet services are configured on an intelligent optical network by a simple point-and-click process using network management. The services are set up on demand automatically versus traditional manual processes that take up to 90 days. Conversely, existing Ethernet services can be modified or removed in the same fashion. Figure 4 represents the actual decrease in service setup time experienced by a service provider as it increased footprint of its intelligent optical network.
In addition to reducing the costs, automation accelerates revenue recognition, provides a competitive differentiator for the service provider, and improves customer satisfaction.
Meanwhile, an inherent feature of having intelligence in the network is that NEs know the entire network including available resources. NE databases are automatically updated when changes occur, which means the information on which service setup decisions are based remains current and accurate, eliminating manual mistakes and errors.
Finally, optical mesh architectures are another byproduct of the intelligent optical network. Mesh topologies are better suited to Ethernet than traditional ring topologies because mesh offers multiple paths through the network. A well designed mesh network will find performance-optimized paths with specific latency, jitter, and loss parameters to meet Ethernet service-level agreements (SLAs) and offer improved reliability and resiliency thanks to its ability to route traffic around multiple failures. In addition, since working service bandwidth is not protected on a 1:1 basis but rather via a pool of shared bandwidth, dramatic cost reductions are achieved. Figure 5 illustrates an example of the savings associated with a real customer network when deploying mesh instead of traditional ring topologies.
Clearly Ethernet has thrived and grown in the enterprise market. It has become the universal choice for cost-optimized switching and transport for all services — particularly the IP applications that dominate the world of telecommunications. Carrier Ethernet is about to experience a significant growth cycle in carrier networks as an end-customer service offering, but most importantly as a next-generation transport infrastructure supporting triple-play, mobile wireless backhaul, and business data service applications. To achieve and maintain this growth, implementations of Carrier Ethernet transport must address the challenges of scalability, reliability, true QoS, end-to-end service management and provisioning, cost-effective interoperability, and transport of legacy technologies such as TDM.
Service provider deployments of Ethernet transport must enhance the end user experience that customers expect. Ethernet can be effectively deployed as a conversion mechanism in access networks, aggregation transport architecture in the metro, wide-area service creation platform at the service edge, and long-haul transport medium across the optical and data cores.
Three key technology pillars have emerged as cornerstones for building Carrier Ethernet optimized infrastructure. G.709 OTN brings low-cost, multiprotocol support with SONET/SDH performance management and monitoring. Pseudowires complement Ethernet/OTN aggregation networks in the metro to enable Ethernet transport over IP/MPLS in the core, and it moves into access networks. Additionally, pseudowires are a highly effective tool for providing local Ethernet transport and Ethernet WAN services. Finally, ITU G.ASON-based intelligent optical networking improves the already resilient OTN with even greater network availability and protection that integrates well with the multipoint/bridging nature of many Ethernet and data VPN services. Its end-to-end provisioning through an intelligent and unified control plane is required for success when enabling Ethernet in traditional transport architectures.
David Parks is senior product marketing manager, data networking, for Ciena (www.ciena.com).