Carriers across the globe are rebuilding their networks on a simplified architecture, migrating individual network segments onto a single converged platform that can adapt to changing market conditions and respond rapidly to customer demands. Until recently, carriers have had to contend with complex and fragmented infrastructures composed of legacy technologies, such as Synchronous Digital Hierarchy (SDH), Frame Relay and Asynchronous Transfer Mode. These networks have become increasingly unwieldy and unresponsive to the bandwidth-intensive applications that are driving much of the revenue potential in the business and residential marketplace.
Carriers are confronted with significant growth opportunities that are creating bandwidth demands in all areas of the network. From the customer premise through the metro access and on to the metro core, traffic is growing at an exponential rate. In the business market, enterprises – including many of the world’s Fortune 1000 organizations – require vital services such as high-speed virtual private networks, business continuity, disaster recovery, continuous data protection, video conferencing, wireless local area networks, and e-commerce channels. These services need to be packaged in a less complex and cheaper communications solution.
The demand for bandwidth is also strong in the residential market where consumers are seeking services ranging from high-speed Internet, video distribution, video on demand, teleworking, video conferencing and online gaming to e-commerce and 3G mobility. As reflected in the business market, residential customers also want these services delivered in simpler, more flexible and more personalized ways.
Bandwidth hunger
A carrier will soon need to provide 100 Mbit/s of bandwidth and services per customer to satisfy demands. The popularity of offerings such as ubiquitous, intelligent Ethernet services for businesses and triple play for consumers represents a great opportunity for carriers. However, growth in carrier profits has not matched the surge in bandwidth and the associated network costs. If a carrier is to remain competitive and respond to customer demands, it must substantially reduce transport-cost-per-bit or face dramatic financial consequences.
European carriers, including British Telecom, Eircom, Fastweb and Neuf Cegetel, understand these financial implications and are leading the industry in a drive for converged transport networks that can deliver all services when and where they are needed. Ultimately, a converged network reduces complexity and, most importantly, lowers capital expenditures and operational costs. It eliminates overlay networks, excess equipment, interfaces and nodes and can draw Internet Protocol (IP), Ethernet and Multiprotocol Label Switching (MPLS) infrastructures onto a single platform. This network simplicity enables carriers to develop architectural flexibility and investment protection as they adopt new tools for dynamic reconfigurability, network scalability and end-to-end planning and management.
The foundation for this network simplicity lies in two fundamental technologies: optical transmission transport and intelligent Ethernet access. Together, these two elements provide an uncommonly intelligent, standards-based combination of optical, Ethernet-aware transport and aggregation technologies that will accommodate more traffic, lower transport cost-per-bit and provide optimal degrees of flexibility and efficiency. Consequently, these technologies are playing a critical role in enabling carriers to migrate from legacy networks and move on to converged infrastructures.
Within the access domain, Ethernet has emerged as the preferred data-link protocol for capacity growth in carrier networks, including next-generation infrastructure for business and residential services. Until recently, Ethernet was considered a “best-effort” service that lacked the carrier-class demarcation and management of its competitors.
However, with the development of intelligent Ethernet technologies this is no longer the case. Carriers can deploy Ethernet solutions that provide advanced remote operations, administration and maintenance (OAM), detailed monitoring of service level agreements (SLAs) and strict demarcation between carrier and customer networks (see figure 1). These services provide the carrier with a previously unattainable understanding of what is happening within their network; what’s more, they do so cost-effectively.
With direct operational and financial benefits, intelligent Ethernet is being widely deployed by carriers both in Europe and the US. In the UK, BT is one of its principal supporters and will be using intelligent Ethernet throughout its new 21st Century Network to provide strict monitoring of activity from the metro core to the customer premise.
Intelligent elements
Within the construction of intelligent Ethernet solutions there are two noteworthy hardware elements that differentiate it from its competitors:
• Field-programmable gate array semiconductors. These devices contain programmable logic components and interconnects that can be programmed both in the manufacturing process and afterwards by the customer. The flexibility to repeatedly re-program these units in the field allows significant CAPEX reduction. Previously, carriers would have had to buy new equipment to meet specific needs and dispose of it when needs changed. With programmable equipment, however, carriers can repurpose the equipment to meet changing customer demands.
• Small form-factor pluggable optics (SFP). This optical transceiver provides the optimum flexibility required for carrier networks that need to carry data across varying distances. One of the benefits to using modern optical SFP transceivers is the digital optical monitoring it supports. This function allows the carrier to monitor real-time parameters of the SFP, such as optical output power, optical input power, temperature, laser bias current and transceiver supply voltage.
Hardware only provides the foundation for intelligent Ethernet; the real intelligence is located in the network and client interfaces. Here there are two primary components, each located at the customer premises, and these are key to extending intelligent Ethernet services:
• The user network interface (UNI), which defines the SLA or service personality, including service policing and definition. Compliant with Metro Ethernet Forum standards, the UNI provides service intelligence for defining committed information rate, peak information rate and burst size per individual ports, priorities or virtual local area networks (VLANs). It uses 802.1p and 802.1q type of service or differentiated services code point to classify traffic into Ethernet virtual circuits (EVCs), VLANs or priority level.
• The network interface device (NID)/network termination equipment (NTE) provides the operations, administration, maintenance (OAM) and provisioning tools needed to support monitoring and testing on both sides of the demarcation point (see figure 2). The NID is 802.1ag/802.3ah-aligned and includes Etherjack Connection Performance Analyzer along with port- or VLAN-level loopbacks to enable a carrier to diagnose faults remotely and verify SLAs conformance for EVCs using RFC-2544 test suites.
Working together, the UNI and NID/NTE deliver a service intelligence that is unparalleled. To ensure this intelligence is reflected throughout the carriers’ infrastructure, intelligent Ethernet is designed to inter-work with a number of core technologies, such as VLANs, Provider Backbone Bridging-Traffic Engineering and Multiprotocol Label Switching (MPLS). This flexibility ensures the carriers’ access network is future-proof and able to adapt to changing market conditions.
Intelligent and transparent
In a next-generation network, however, it is not only Ethernet that possesses service-level intelligence and protocol transparency. Optical transport transmission built on wavelength division multiplexing (WDM) provides the ideal network partner. It can support any bit rate and protocol integrating both legacy and next-generation, intelligent carrier-class Ethernet/IP services across all tiers of the network. It offers low-cost transport for all applications and services, scales easily in terms of capacity and reach and provides rapid protection against any fibre plant failure.
Optical WDM is the ideal, multi-service integration layer for modern networks – rendering them more manageable, more efficient and more capable of delivering the tremendous volumes of bandwidth required for the emerging class of business and consumer services.
Many of these services have unpredictable bandwidth demands. For example, organizations seeking to harness grid computing, virtualization and server consolidation will have intermittent periods of intense bandwidth activity. Unless the organizations are using dedicated private networks it is vital for carriers to maximize bandwidth capacity and eliminate redundancy.
Today’s leading WDM systems can deliver this service with fully automated reconfigurable optical add/drop multiplexer (ROADM) capabilities that provide an unrestricted use of wavelengths at any node. Two- and multi-degree ROADMs are the key technologies that deliver this.
If a carrier seeks flexibility in its infrastructure, ROADMs provide the ability to establish circuits and light paths across the optical network dynamically via simple requests. The ability to switch light paths to multiple network ports dynamically is a feature of ROADMs that contain Wavelength Selective Switch (WSS) functionality, allowing any-port connectivity across supported topologies. WSS is a significant advance in ROADM technology and will have great impact on reducing OAM costs.
A WSS provides switching and wavelength equalization. Responding to software commands, it performs all mux, demux, optical switching and attenuation functions. Initially, ROADMs featured Dynamic Channel Equalizers that performed only block-or-pass functions. Now, with WSS integration, ROADMs have the ability to fully switch wavelengths. Figure 3 shows the interconnection of two rings with a multi-degree ROADM using WSS technology, enabling reconfigurable services and bandwidth-on-demand.
ROADMs to the future
Modern ROADM architectures such as this offer three distinct advantages:
• The ability to add two-degree or multi-degree ROADMs when and where they are needed. Without having to plan for unknown traffic demands, ROADMs eliminate stranded equipment and drive CAPEX and OAM costs down.
• The option to perform in-field upgrades of ROADMs from two-degree to multi-degree.
• The significant lengthening of the lifecycle of the hardware with the use of software automation.
A seamless integration of ubiquitous, intelligent Ethernet combined with aggregation and switching into WDM systems provides the foundation for true end-to-end networking. An Optical+Ethernet approach obviates the discussion about which technology best suits which application, and consequently eliminates investment risks (see figure 4).
The future for next-generation networks can be divided into three distinct phases: convergence, then dynamic utilization of resources and, finally, virtualization. In phase 1, we are witnessing the convergence of data, voice and video applications as network operators seek the ease of use, integration and consistent functionality of a unified network system for optimized Ethernet service delivery over optical networks. Phase 2 will bring more efficient and flexible use of resources by the network, with increasing deployment of control plane and multi-degree ROADM functionality. In phase 3, applications and services will be “virtualized”, with network and applications tied more closely together to ensure automatic resource provisioning as necessitated by the application.
Optical and Ethernet technologies provides carriers with powerful and flexible tools to meet the challenges they will face over the course of this evolution, including how to handle traffic growth while controlling costs, how to provide scalability and service quality to drive maximum revenue and how to encounter emerging drivers in terms of cost and technological advancement. These attributes make integrated optical and Ethernet the ideal, long-term network solution.