White Paper
Executive Summary
Customer demand for increased bandwidth and new network services is growing rapidly and steadily. Industry analysts such as IDC, Infonetics Research, and Communications Industry Researchers, Inc. project that Metro Ethernet service revenue alone will surpass US$630 million by 2006 in the United States and $9.7 billion worldwide, with $7.9 billion coming from the burgeoning Asia Pacific market. Service providers that have historically relied on voice-grade networks are increasingly using IP to help enable new bandwidth-intensive services, such as Metro Ethernet and high-speed optical. To support these services, many service providers have built overlay networks on top of existing time-division multiplexing (TDM) environments.
While this overlay approach solves the short-term problem of bringing a new service to market, it presents several long-term challenges. A network infrastructure that incorporates multiple service rings is complex, inflexible, and difficult to scale. Such a network is also often extremely costly to maintain because it requires a variety of technologies, management and provisioning tools, and staff expertise. These costs increase over time, because service providers often must limit existing overlay networks and build new ones to meet ever-growing customer demand.
Fortunately, service providers have another option: Instead of building new overlay networks to support each new service, they can evolve their current networks to a multiservice optical infrastructure. Unlike a traditional network, a multiservice optical network converges, or merges, all service rings and network layers into a single optical infrastructure. This next-generation infrastructure supports any type of network traffic and any transport technology, helping to enable streamlined end-to-end provisioning of new services, regardless of customer premises equipment (CPE) or network interfaces. In this evolved statistical-multiplexing environment, service providers can maximize service density and capacity, reduce network complexity, and help enable a long-lasting foundation for supporting higher-margin services.
Market Changes Promote New Opportunities
Service providers are beginning to see a dramatic shift in customer needs and expectations. As enterprise usage of new services grows, demand for bandwidth is also rising dramatically - and both show no signs of abating. Several factors are influencing demand:
• Growth in high-bandwidth services and applications - Emerging from an economic downturn, enterprise companies are more focused than ever before on improving employee productivity and operational efficiency. As enterprises deploy new applications and productivity tools, they increasingly demand higher-bandwidth connections for voice, data, video, and storage applications.
• Expanding use of storage solutions - With the heightened awareness of potential terrorist attacks and fast-spreading viruses, many enterprises have placed an increased emphasis on the need for geographically dispersed data stores. The need to back up enormous amounts of data and synchronize databases across a wide area has triggered new demands for storage-area network (SAN) solutions and the bandwidth to support them.
• Growing consumer broadband usage - Broadband Internet service has already made substantial inroads in the consumer market. With new consumer technologies and expanded high-bandwidth content expected to emerge over the next several years, service providers expect to see consumer bandwidth requirements continue to grow.
Although this demand for more bandwidth and new services may present short-term challenges, it is a tremendous long-term business opportunity for service providers. For business Ethernet services provide scalable bandwidth in flexible increments, with simplified management and faster, lower-cost provisioning. Metro Ethernet complements service provider portfolios by enabling the delivery of data-optimized, high-bandwidth services that address a significant new market opportunity. Yankee Group predicts that the market for metropolitan-area network (MAN) and WAN services will grow tenfold in the next four years, from $66 million in 2003 to $651 million in 2008.
To capitalize on these opportunities and remain competitive in the long term, many traditional service providers are offering new, value-added services. Doing so, however, requires creating the necessary infrastructure to support them.
Evolving the Service Network
As service providers begin to capitalize on new revenue opportunities and improve service to their existing customer base, many are transitioning from TDM environments to statistical multiplexing, or packet-based IP environments. IP networks are optimized for the "bursty" nature of data traffic, but have historically been provisioned and managed very differently than traditional TDM infrastructures. For example, traditional TDM services such as voice circuits, 1.5Mb/sec. (DS1), and 44Mb/sec. (DS3) services have be supported and managed from a central office Class 5 switch or from SONET optical transmission equipment, whereas IP services have been managed with a very different interface, directly from the console ports of switches, routers, and IP gateways located more to the edge of the SP network. As a result, many service providers have built overlay networks to support new IP service offerings, such as Metro Ethernet, VPNs, high-speed optical, videoconferencing, and IP telephony.
Service providers can now take advantage of converged optical network technology to eliminate the need for multiple overlay networks and their associated costs and complexities. A multiservice optical network converges multiple linear services into a single optical ring, and combines Layer 1, Layer 2, and Layer 3 technology into a unified network platform. Service providers enjoy a smaller footprint, fewer network elements, and convergence of the physical interfaces required to support new IP services onto a single platform, which can interface with virtually any type of customer premises device (see Figures 1 and 2).
Figure 1
Multiple Networks for Multiple Services
Figure 2
Converged Transport Network
By evolving to a converged optical network, service providers can build a more efficient, robust, long-term foundation for meeting the full range of growing customer needs. This foundation helps enable continuous deployment of high-margin, revenue-generating services, while optimizing both customer-facing and back-office business and network operations.
Benefits of an Evolved Network Infrastructure
By evolving to a multiservice optical network, service providers can dramatically improve business efficiency, profitability, and long-term scalability. An evolved multiservice optical network enhances:
• Service variety by supporting both current and new types of services, creating a foundation that will support new applications, service packages, and revenue streams
• Service density by maximizing the amount of traffic in available bandwidth
• Service velocity by helping enable faster deployment of both current and new services to meet customer demand
• Service capacity by helping to ensure scalability across all platforms, interfaces, and distances for meeting future needs
• Service efficiency, dramatically reducing the costs of provisioning and managing network services
Service Variety
For many traditional voice carriers, expanding service offerings may mean simply offering more WAN services. With an evolved multiservice optical infrastructure, the potential for deploying new services is virtually unlimited. Service providers add line cards to support each new offering on a multiservice optical platform, and they can begin provisioning the service almost immediately. The management interfaces, tools, and processes are virtually the same, regardless of the type of service being deployed. After the line card for a specific service is deployed, the network can immediately communicate with equipment at the customer office in the customer's native interface. As a result, service providers can cost-effectively bring the full range of wavelength and subwavelength services to edge access points in office buildings, campuses, and residences.
In addition, by delivering all services over IP, service providers can use previously inaccessible Layer 2 and Layer 3 intelligence to gain unprecedented granular control of these services. Providers can offer customized bandwidth, bandwidth on demand, and customized quality of service (QoS), helping them to craft more competitive offerings. Service providers can tailor services for different times of day, adjust services at the customer's request, and handle different types of traffic with different degrees of latency. All of these options can be formalized in service-level agreements (SLAs), which provide additional revenues for enhanced levels of service.
A converged multiservice network offers the same TDM-grade reliability customers expect. Technologies such as Resilient Packet Ring (RPR) help service providers effectively provision Ethernet over a SONET/SDH network, while continuing to support highly reliable TDM services. This enables service providers to offer a wide variety of services over the same infrastructure. Services that were previously segregated (such as voice and Ethernet, or Ethernet and video) can become integrated, offering customers advanced productivity benefits and fundamentally changing the nature (and profitability) of many existing services.
Expanding Optical Service
By evolving to a converged network that supports dense wavelength-division multiplexing (DWDM), service providers gain a completely transparent, protocol-independent optical provisioning technology. Such a network can carry any type of transmission, and any transport or storage protocol. In the past, service providers used DWDM primarily to deliver virtual fiber applications to customers, which was prohibitively expensive. A converged optical core provides service providers with a lasting foundation for meeting customers' bandwidth needs for the next decade and beyond.
A converged wavelength and subwavelength network also helps enable service providers to deploy high-bandwidth optical as easily as any other service. For example, multiservice optical platforms from Cisco Systems® support reconfigurable optical add-drop multiplexing (ROADM), which allows service providers to deploy optical services end to end using a simple point-and-click interface - rather than using complex optical wavelength tools and highly specialized optical staff. The Cisco® ROADM solution provides any-to-any connectivity without the need to predetermine the optical path. This simplifies network planning and engineering and eliminates the cost of storing multiple spare parts for specific drop locations.
Service Density
An evolved optical network allows service providers to carry increased traffic over existing infrastructure and reduce cost per bit throughout the network. A single optical platform with Cisco ROADM capabilities, for example, can support up to 32 redundant wavelengths on a single card, providing high levels of density and scalability. An evolved network can also use advanced technologies, such as multirate optical line cards with Cisco Small Form-Factor Pluggable (SFP) optics that help enable higher port density and enhanced efficiency. With SFP optics, a single card can support all the services that multiple cards would support in a traditional platform, reducing capital costs and opening shelf slots for other services.
Service density can be expanded even further using service aggregation to carry multiple service types per wavelength and per fiber for efficient use of bandwidth. For example, a Cisco optical platform might support 10 ports of ESCON, four ports of Fibre Channel, and two ports of Gigabit Ethernet over a single wavelength. This results in more efficient use of wavelength and virtually unlimited fiber capacity.
Adding Intelligence to the Service Network
Service providers also help enable higher service density by converging multiple layers over a single infrastructure. Historically, many service provider data networks have offered little intelligence or transparency in the transport of different types of data traffic. An evolved network core can use the statistical multiplexing capabilities of IP and Layer 2 and Layer 3 to incorporate traffic queuing, QoS, and other intelligent services. This helps service providers give customers the amount of bandwidth they need at any given time, while also conserving excess bandwidth to offer more services and attract more revenue, without deploying new infrastructure.
In a traditional multiplexing environment, service providers can only provision bandwidth in discreet chunks, which may not match a customer's real needs or usage patterns. For example, a customer with a DS-1 connection might require a higher-bandwidth connection one day each week. A service provider with TDM ports may have no choice but to upgrade that customer to a DS-3 connection - despite the fact that six days a week, the majority of that bandwidth is never used. This results in the customer paying a higher price for a largely unneeded service, while the service provider loses valuable bandwidth that could otherwise be devoted to other revenue-generating customers. In a statistical-multiplexing environment, the service provider can retain the existing connectivity and increase the bandwidth connection only when the customer needs it - for a more competitive offering and a more efficient use of valuable network bandwidth.
Service Velocity
In the past, service providers had extensive planning timelines for deploying a new service, and once the service was deployed, the network could last for many years. Today, the planning horizon has all but vanished. Customers have more demands and more options, and service providers are working to bring the newest, highest-margin, and most competitive offerings to market as quickly as possible.
Unlike a traditional network, service providers with an evolved multiservice network are not limited by the type of infrastructure in the central office, the access point, or the customer premises. An evolved network can accommodate any new type of service easily and in the preferred format, without requiring customers to adapt to the service provider's infrastructure. Using technologies such as Cisco SFP optics, service providers can easily change bandwidth speeds and frequencies of optical connections in a few seconds. Instead of needing a highly specialized optical staff to deploy a new line card or platform, service providers can simply exchange the SFP on the existing line card.
An evolved network also helps service providers to provision any service - voice, Ethernet, and even optical - end to end through a unified management interface. Instead of asking customers to wait days or weeks for a service change to be processed, service providers can make changes in minutes, increasing or decreasing the level of services with a click of a mouse. Higher service density also means more capacity and more efficient use of the network. Customer requests for more bandwidth can be handled through a network management tool, as opposed to a service call. The end result is faster, more responsive, and more cost-effective service.
Service Capacity
By moving to a multiservice optical network in which all services are carried over fiber, service providers can help ensure that the network continually meets growing customer demands. Fiber can theoretically carry a limitless amount of traffic, which places service providers in an ideal position to take advantage of new market opportunities. In a traditional network, technologies are segregated by distance, with dedicated platforms for access, metro, regional, and long-haul services. An evolved optical network converges all of these technologies, and can effectively transport network traffic.
By converging multiple network layers and employing IP intelligence, service providers can also maximize bandwidth in any service ring. At the central office, an evolved network supports multiple TDM and data services and integrated DWDM transport over a single infrastructure. This eliminates the need to provision multishelf port interconnections that reduce port capacity while increasing costs. Service providers are never limited by any customer's infrastructure or network format because an evolved optical network operates independently of the customer's LAN interfaces.
Cost Reduction
An evolved network that converges network layers, services, and network elements can dramatically reduce a service provider's capital costs, operational costs, and total cost of ownership. In capital costs alone, a multiservice optical platform reduces:
• Short-term expenses for deploying new services - Instead of investing in an entirely new overlay network, as well as all the training and resources required to manage it, service providers can simply upgrade the functionality of existing infrastructure.
• Network equipment footprint - Consolidating network layers eliminates the need for many separate network elements, such as Layer 3 routers at customer premises and channel extenders, which can result in significant savings. Using technologies such as virtual concatenation (VCAT) and link capacity adjustment scheme (LCAS), service providers can also access any individual circuit on a service ring, reducing the amount of multiplexing and demultiplexing equipment in the network. A reduced footprint also means reduced heating, ventilation, and cooling costs.
• New fiber investment - Using next-generation technology to help ensure the most efficient use of existing fiber infrastructure, service providers can delay investment in new fiber plants, potentially indefinitely.
• Spares inventory - A service provider with multiple overlay networks must possess numerous spare components for each network. A converged multiservice network requires a much smaller number of spares. For example, a service provider using Cisco ROADM technology can keep a single optical card spare capable of supporting 32 different wavelengths, as opposed to storing separate filter cards for all 32 interfaces.
• Future investment - A traditional TDM platform or data-only overlay network has limited scalability and capabilities, and is very expensive to expand. Alternatively, a hybrid Layer 1-3 network offers a comprehensive suite of services and a combination of transport options that eliminates the need to continually cap existing networks and deploy new ones.
Operational Cost Savings
An overlay approach carries long-term operational costs that can minimize the new service deployment investment, making the operational cost savings of an evolved multiservice network even more substantial. Service providers have an installed employee base accustomed to maintaining a specific network, using a specific set of tools and skills. Each new overlay network requires new equipment, new installation procedures, new card stocks, and new management interfaces. Often, this translates into separate workforces managing each new service. As customer needs change, an overlay approach helps ensure increased network complexity and the continued growth of resource requirements to support the service network.
With an evolved multiservice optical network, service providers can add capabilities to their existing network without needing to substantially change equipment, management interfaces, or employee skill sets. In this environment, service providers can add Ethernet, SAN, or wavelength services by plugging a new card into the existing shelf. Employees provision the new service in the same way they would provision a TDM circuit. Service providers benefit from reduced staffing and training requirements, improved productivity, and most importantly, reduced service calls, because nearly all tasks can be conducted virtually, over a computer interface. Billing is also more efficient, because only one management interface must be queried to generate bills for the full range of services.
Which Networks Can Benefit From Evolution?
Network evolution is not the answer for all service providers in all situations. Asking specific questions can help identify whether your network is a good candidate.
• Is the network at or near capacity?
• Is it difficult to respond to new service requests? Are customers choosing to work with competitors who can offer a wider range of services and more competitively priced service bundles?
• Is a dependence on traditional SONET/SDH technologies limiting the services that can be deployed?
• Is the addition of overlay networks required to respond to new service needs?
• Is the network built primarily as redundant, end-to-end linear circuits? Is it composed of back-to-back network elements in a multiple-ring configuration?
• Does the network require a large footprint of multiplexing and demultiplexing equipment to enable the delivery of specific circuits to specific customers?
• Is scarce network bandwidth being handled with maximum efficiency? Is bandwidth available to customers "on-demand?" Are limited interface options forcing customers to pay for more bandwidth than they need?
• Is it possible to respond to new service requests rapidly?
• Is the network flexible and scalable enough to grow one service with respect to another as an individual customers' needs change?
• Is there a dependence on separate equipment, management interfaces, spare inventories, and staff to maintain multiple service networks?
Converged Network Deployment Scenarios
Network evolution is a proven strategy for streamlining network operations, maximizing bandwidth, and creating new, high-margin services. Many service providers worldwide have already benefited from this strategy, and Cisco has developed extensive best practices and blueprints for achieving a successful evolution. The following examples illustrate some of the advantages associated with different types of network evolution.
Traditional Private-Line Network
In this example (Figure 3), the service provider uses a traditional private-line overlay network to provide data connectivity for 40 distinct sites, representing 13 customers. This network requires seven aggregation nodes and 26 separate circuits - a two-to-one ratio for each customer.
Figure 3
Private-Line Network
This network of point-to-point circuits is inherently inefficient for providing bursty data services, because many of these customers are most likely not using all of the bandwidth allocated to them in each dedicated circuit. In addition to presenting a more expensive, less efficient, and less attractive offering to customers, this network is also more costly for service providers to provision and maintain. Figure 3 represents customers with only a few sites each, and a single service offering. As customers add more sites or demand new services, more dedicated circuits are required, making the network more complex and more resource-intensive to manage.
Shared-Capacity Network
In this example (Figure 4), the service provider is able to serve all 13 customers and 40 sites with a single circuit. The total bandwidth is shared; however, each customer's data traffic is logically separated to help ensure security.
Figure 4
Shared-Capacity Network
In this scenario, all 13 customers are served as a single VCAT group, so the service provider can use VCAT and LCAS to adjust bandwidth as needed to each customer and each site. The result is a network that more efficiently maps available bandwidth to each customer's actual needs. Instead of paying for a dedicated circuit and bandwidth that they do not need, customers receive a lower-cost, more flexible solution. In addition to being able to offer a more competitive solution, service providers gain increased service capacity because they can serve more customers with less bandwidth.
A single, shared-capacity network also allows for more simplified management than a mesh of point-to-point circuits. Instead of layering point-to-point data networks on top of an existing TDM ring, the service provider can use a single multiservice ring and a single network element type, and easily provision a variety of revenue-generating services.
Evolved, Dedicated Network
A shared-capacity network is ideal for many customers, especially small and medium-sized businesses (SMBs). However, it will not meet the needs of every customer. For example, a financial institution or healthcare organization may be required by law to have a separate network. Or, some customers may have traffic requirements that make a shared medium impractical. For these types of customers, service providers can still provision dedicated domains in a much more efficient manner than in a traditional mesh architecture (see Figure 5).
Figure 5
Dedicated-Capacity Network
In Figure 5, the network includes one dedicated circuit for each customer. Unlike a traditional point-to-point network, all customer sites share a single circuit - instead of requiring several dedicated circuits to connect each location. In this scenario, the service provider is also using RPR to help ensure network availability, rather than deploying redundant physical circuits for each customer. As a result, even having13 customers with their own domains, the circuit-to-customer ratio is one-to-one, instead of two-to-one in a traditional private-line network - a 50 percent reduction in the network footprint that must be managed.
Evolution Strategies
Service providers have concerns about evolving their current network infrastructures, especially when these networks continue to serve customers and function as they were designed. The long-term scalability and revenue potential of an evolved network coupled with the high costs of maintaining an overlay approach will lead many service providers to decide that evolution is the best strategy. With proper planning, service providers can evolve their networks in a cost-effective manner and rapidly begin benefiting from multiservice optical technologies.
Many service providers take an incremental approach, beginning by moving current core equipment outward, and redeploying it at the network edge. Instead of considering the previous network a lost investment, these assets are instead used as the foundation of an exponentially more robust service network. Moving previous core elements to the edge allows the network to become more capable of growing, scaling, and generating continuous return on investment (ROI) than if the service provider had simply deployed a new overlay service.
If a service provider has Cisco equipment that cannot be redeployed at the edge, Cisco offers a wide range of competitive trade-in programs to help service providers accomplish their strategic objectives without incurring significant costs.
Conclusion
The market for voice, data, and other network services is changing rapidly as customer processing power and bandwidth requirements continue to grow. Service providers have many options for responding to these changes. However, no service provider wants to continuously reconfigure networks to support the latest service and remain competitive, while also trying to limit new capital costs.
Forward-looking service providers are choosing to invest in a one-time upgrade that boosts the capacity, density, service velocity, and service variety of their networks. These service providers are evolving traditional SONET/SDH infrastructures to next-generation multiservice optical networks. Through evolution, these service providers are achieving a more efficient, longer-lasting network infrastructure that will continue to serve their customers and generate ROI into the future.