Metro Solutions / SPECIAL REPORTS
An integrated-services platform that supports the migration from commodity bandwidth to high-margin IP and WDM-based services.
RAJIVE DHAR, Atoga Systems
Optical technology and rapid deployment of bandwidth-hungry applications are redefining the metro network of the future. Content distribution is becoming vital as more businesses rely on service providers for content hosting, data mirroring, and data backup. While these services have caused service providers to redefine their business models and push for network architectures offering the flexibility to migrate to newer services faster, challenges remain on many fronts. Among these challenges are anticipating demand for services, providing the right bandwidth for customer applications, and picking technologies that provide the highest returns on investment.
The opportunities that resulted from deregulation produced a very chaotic telecommunications market. Profitability was sacrificed at the cost of grabbing the maximum footprint, even if serving that footprint increased a service provider's cost by a factor of 10.
Today, the situation is different. Service providers want high margin, value-added applications, delivered in the most reliable manner at the lowest possible cost, to generate positive cash flow from operations.
How does a service provider achieve these divergent goals? While providers universally acknowledge the future communications infrastructure will be all-optical to achieve these goals, individual service providers have different views on deployment.
Whether it is achieving the lowest cost-per-bit solution, solving the fiber exhaust problem, or enabling high-margin value-added services, providers need an integrated-services platform that supports today's legacy time-division multiplexing (TDM) bandwidth services and provides a migration path to margin-rich Internet Protocol (IP)-based services as well as next-generation dynamic wavelength services.
Key attributes of this integrated-services platform include optical scalability, fast provisioning, multilevel protection, carrier-class stability and performance, efficient quality of service (QoS), and a simplified network-management system.
A major bottleneck in today's network architecture occurs when traffic must be converted from optical to electronic form for processing, then back to optical before it is sent along the network. This optical-electrical-optical (OEO) conversion is both costly and time-consuming. Service providers estimate that their capital and operating costs could be reduced by as much as 50% by eliminating optical-electrical (OE) conversions. The (OEO) conversions can also make the network unsuitable for the speed-intensive applications that service providers must roll out to migrate to high-margin services.
By incorporating tunable lasers, the integrated-services platform intelligently grooms traffic onto wavelengths within the network, eliminating unnecessary packet processing at intermediate nodes within the network. Tunable lasers also ensure that service providers are not bound by bandwidth since additional channels by changing the laser frequency.
In contrast, capacity increases today require the lighting of additional dark fiber. That involves securing rights of ways to lay a conduit, or if a conduit is already available, the purchase of dark fiber and transmission equipment. Significant engineering is then required to configure and integrate the additional crossconnects and add/drop multiplexers into an existing network.Service providers are eager to activate services faster in the metro. However, a single provisioning event can involve configuring multiple network layers and boxes, making the process both lengthy and expensive.
Today, no standards govern how provisioning requests are communicated among IP, SONET, and WDM devices. Equipment at the IP layer exchanges information with other IP equipment in the network, and equipment at the SONET layer exchanges information with other SONET equipment in the network. But there is no exchange of configuration information between IP and SONET equipment. This communication gap forces the equipment at each layer to guess at the traffic loads to be processed at the other layer, and the fallout is gross inefficiency in service provisioning.
Integrating all the layers of the network architecture vastly simplifies this process, since all adjacent-layer communications are within a single box. Provisioning becomes a seamless event across the network layers, paving the way for more aggressive delivery of high-margin services in the metro (see Figure).
Service providers' infrastructures must be reliable and resilient to faults. While routing protocols afford a measure of fault protection to IP traffic, restoration is not fast enough. In a worst-case scenario, network restoration can take a few minutes.
The integrated-services platform combines SONET's fault detection and isolation capabilities with the optical protection provided by tunable lasers. The restoration takes only a few milliseconds-within the 50-msec restoration offered by SONET. This multilevel protection also ensures that network failures are healed before higher-layer protocols, or users, detect a fault.
High-margin services require appropriate service guarantees. The lack of mature traffic engineering features in today's networks often results in over-provisioning. These overprovisioning costs can run as high as 30%-40% of total network costs.
A "nailed-up" TDM circuit implies a fixed connection that is dedicated for the benefit of a specific customer. While this circuit is always available to that customer, it lies wasted if not used. This wasted bandwidth creates a pricing inefficiency: The customer can't justify paying for the nailed-up bandwidth and the service provider has no way to offer a flexible bandwidth service.
Advanced traffic engineering technologies can enable a prioritized QoS architecture, which is critical to bringing service guarantees to packet-based traffic. Service providers can then manage applications with respect to delay, jitter, and loss to differentiate their service offerings.
Managing a metro network spread across miles on an element-by-element basis is labor- and capital-intensive. Reductions in capital and operating expenses are fundamental to a service provider's profitability. Studies indicate that operating expense represents about 30%-50% of network costs.
An integrated-services platform reduces the number of elements that need to be managed across a network. A single platform is sufficient where previously three separate boxes were needed to deliver IP and TDM services.
A centralized policy-based management system further simplifies network management by removing the need for constant management of the network elements, providing additional savings for the service provider. With a centralized policy server, network usage policies can be automatically applied to any network elements. These usage policies are determined and enforced based on customer-specific service-level agreements. This directory-based policy management also provides visibility into the network and its traffic characteristics ensuring optimal use of network resources.
Newer high-margin services, such as bandwidth trading and IP transit, necessitate information exchange at all levels within the network. Support for these services requires protocols that enable efficient allocation of bandwidth among users in a flexible and dynamic manner. Routing protocols such as open shortest path first, intermediate system to intermediate system, Multiprotocol Label Switching, differentiated services, and border gateway protocol are key enablers for these services. These protocols provide enhanced network scalability, quality-of-service, and class-of-service controls, and standards-based interoperability-features necessary to support migration from TDM-based commodity bandwidth services to high-margin IP services.
To compete in today's market, carriers need to deploy IP and TDM services such as video on demand, online gaming, content hosting, data mirroring, and data backup as well as traditional voice services.
Each of these services has a diverse set of requirements, from cost structure and growth pattern to technology and deployment timeline. In the past, separate network equipment was necessary to support the deployment of these services; today, an integrated architecture can reduce the complexity and expense of deploying these services. With an integrated architecture, carriers deploy high-margin "services" rather than "configure" boxes.
Rajive Dhar is director of solutions marketing at Atoga Systems (Fremont, CA). Dhar can be reached at [email protected].