Optical broadband services key to last mile

Last-mile access often represents a significant challenge for carrier deployment of broadband services, due to the costs of fiberization. Moreover, this cost is dominated by charges that are largely unaddressable, such as right-of-way and labor. As a result, most broadband service deployments have relied upon moving the aggregation point closer to the edge and utilizing copper technologies for the last mile.

Recently, however, a confluence of business and technology developments has made the case for fiberization of the last mile to enterprise customers more compelling. These developments have included a declaration of victory for Ethernet as the local-area-network (LAN) protocol of choice and its subsequent robustness enhancements for wide-area-network (WAN) connectivity.

The almost overnight development of a storage connectivity requirement to meet business continuity and disaster recovery guidelines-and several innovations in the alphabet soup of telecom standards such as GFP, VCAT and LCAS-have made the core network much more data friendly.

We all know the refrain by now: Ethernet has won the LAN with an estimated 98 percent of total port shipments expected this year. Multiple robustness enhancements, however, make Ethernet a contender as the WAN protocol of choice. These enhancements include transparent domain identifiers for security and scale; broadcast control mechanisms like MAC caches; and protection schemes like split multilink trunking and resilient packet ring. Combined with the scale of optics, "optical Ethernet" provides a simplified WAN networking scheme that eliminates expensive adaptation and the complexity and cost of multiple protocols.

Enterprises are intrigued by the operational simplicity optical Ethernet delivers but carriers have been less eager to jump on the bandwagon, fearing the cannibalization of lucrative legacy services such as private lines and frame relay. Joint studies undertaken collaboratively between Nortel Networks and leading enterprises, however, indicate that those fears may be unfounded.

In fact, optical Ethernet services represent incremental revenue opportunities for service providers because they can deliver quantifiable savings to the overall enterprise IT budget, not just for connectivity. Today, less than 10 percent of a typical IT budget is allocated for telecom services. An optical Ethernet service modernization offers the promise of substantial savings to the remaining 90 percent through centralization of computing resources and elimination of duplicate hardware, software and support staff. One study conducted with a Fortune 1000 company found that increasing the telecom component of the overall IT budget to 13 percent produced a total budget reduction of 7 percent, including a 33 percent operations cost reduction, through a dramatic simplification of network complexity and maintenance. This provides a compelling business case for the enterprise service provider, along with a 30 percent increase in potential revenue: a compelling growth opportunity in a challenging marketplace.

Storage of data is big business today. But recognition of the need to safeguard it and provide an effective means of rapid recovery is a recent phenomenon driven by world events and well-publicized outages. Indeed, estimates show only 3 percent of current disaster recovery plans would succeed in one day and only a small fraction, 15 percent, would achieve a full recovery at all, according to a study by Compass America. Regulators, particularly in the United States and the European Union, are responding with guidelines intended to safeguard companies and individuals from the loss of records. The net result is a need for geographically diverse storage of high volumes of data. This represents perhaps the biggest incremental revenue opportunity for service providers over the next few years.

Storage applications can be thought of simplistically as being either real-time (that is, synchronous)-typically used for transactional data-or asynchronous, in which case the archiving of information is usually done during off hours. Both types are highly sensitive to packet loss, which forces a reinitiation of the data transfer. The protocol of choice for storage applications is, therefore, Fibre Channel, which is designed for lossless packet environments with predictable data volumes.

Ethernet and storage can both be adapted through a single mapping technique for all services, Generic Framing Procedure (GFP), recently standardized by the ITU as ITU-T G.7041. GFP provides a consistent mapping technique for multiple service types (storage, Ethernet, etc.) that is Layer 2-plus agnostic (by preserving MAC addressing, for example), with a deterministic low overhead tax. It is an interoperable global standard that provides a clean mapping of broadband traffic to Sonet/SDH (or OTN) frames. GFP allows service providers to maintain consistent network management and control systems for these attractive new services as well as existing ones.

GFP is complemented by virtual concatenation (VCAT) and the link capacity adjustment scheme (LCAS), both also standardized by ITU as G.707 and G.7042, respectively. VCAT provides a mechanism that allows GFP-mapped signals to occupy "virtual" or noncontiguous Sonet/SDH slots. These are carried independently through the network and reassembled at the terminating end. This setup provides great flexibility in the deployment of high-bandwidth services over an existing congested Sonet/ SDH network. It effectively handles future service churn by separating the port speed of the incoming signal from the actual rate of service. One can increase the rate of speed with minimal impact on the enterprise or carrier, since neither the port nor the underlying infrastructure requires upgrading (if sufficient total capacity already exists).

Internal Nortel Networks analysis indicates combining GFP and VCAT yields an approximate 40 percent increase in network efficiency over a typical mix of wideband and broadband traffic and a possible increase in provisioning cycles of more than 60 percent. Provisioning enhancements are based on the number of circuit adds/deletes that can occur before reaching a prescribed level of system fill and the network's ability to address churn. VCAT provides a dramatic improvement based on its ability to work around fragmented or stranded bandwidth.

LCAS provides a mechanism to assign this additional bandwidth on a perservice basis, hitless and in real-time. In addition to being valuable for planned service changes, this capability also enables differentiated classes of service with guaranteed throughputs of either the full rate or a reduced subrate in the event of network degradation.

Further, these Layer 1 techniques are augmented by packet techniques that derive the last ounce of efficiency and reliability from the fiber, and realize the cost "sweet spot" required to make optical Ethernet services available to the masses.

Together, these standards-based technologies allow service providers to leverage their existing Sonet/SDH infrastructure in the introduction of new service types, creating a multiservice network that is thereby efficient not only for Layer 1 services but also transparent for mapped Layer 2-plus services. In doing so, the end-to-end service delivery cost is reduced. This efficiency, combined with the incremental revenue opportunities that storage and optical Ethernet services present, creates the necessary business case for fiberization and service modernization.

For additional information and updates on Nortel's alliances and collaborations with other companies, go to www.nortelnetworks.com.

Other related alliances and forum information can be found at these sites:

• Ethernet in the First Mile Alliance: www.efmalliance.org

• Metro Ethernet Forum: www. metroethernetforum.org

• MPLS Forum/Frame Relay Forum Alliance: www.mplsforum.org

• Optical Internetworking Forum: www.oiforum.com

Rob Keates is a senior marketing manager in optical networks at Nortel Networks (Alpharetta Ga.). He can be reached at rkeates@nortelnetworks.com.

http://www.eet.com