With new technology enabling the provisioning of any type of service over any type of access network to any type of user, maybe it's time for the telecommunications industry to start treating value-add in data the same way it treats it in voice -- as feature sets rather than as a hodgepodge of virtual and managed services.
"We're using the wrong terminology in an industry where people are accustomed to selling features from a Class 5 switch," suggests John Holobinko, president and CEO of Aplion Networks, a provider of next-generation service provisioning technology. "We don't call them managed voice services, and we're at a point now where we shouldn't be complicating the idea of data feature provisioning by calling them managed data services."
While Holobinko is referring to the highly flexible, point-and-click capabilities of the Aplion platform, his comments also apply to the larger reality of what's happening in the push to create new revenue streams. As Raj Sharma, director of engineering network architectures at Luminous Networks, puts it, "Carriers recognize they have to create a stable pricing model for these new services."
This year's SUPERCOMM demonstrated that the highly condensed multiprotocol modules available for core, edge and premises deployments is erasing the division between Layer 2 and Layer 3 services, where what can be offered depends on whether the service is going over the network as a TDM frame, a cell or a packet. True, the service provider must sort through a mind-boggling array of technology to find just the right combination of platforms to transform its legacy network into the multiprotocol provisioning engine of the future, but no matter what type of access and metro backbone, it can be done.
Examples too numerous to mention were present on SUPERCOMM's convention floor. A review of some of the innovations making their first appearance provides a good perspective on the tools carriers now have at their disposal to make feature-rich data service provisioning a mainstay of their networking capabilities without having to replace existing infrastructures.
In the case of Luminous, the new wrinkle was introduction of a customer premises platform supporting multiservice access over resilient packet ring networks. The company's PacketWave E-Series optical access platform, positioned in multitenant units, campus networks or business parks, gives carriers the means to provision the full range of differentiated Ethernet services while supporting TDM voice and circuit services over the same distribution facilities, Sharma says. Luminous, which now uses Ethernet as the physical layer in its PacketWave products, soon will create a SONET transport layer, making it possible to piggyback the RPR access transport onto existing rings, he adds.
Taking another approach to facilitating integration of multiple services, ADC Telecommunications has come up with what it calls the iAN (intelligent Access Network) gateway, which is designed to enable delivery of converged data and voice services over T1 lines while providing highly granular service provisioning capabilities in the IP and Ethernet domains for DSL and other access platforms as well as T1. Where ADSL is concerned, the technology "allows you to add additional IP services on top of the plain vanilla without having to sell the customer higher bandwidth tiers," notes Sayeed Rashid, director of marketing for broadband infrastructure and access at ADC.
ADC's approach to making this possible involves use of a service gateway positioned at the central office in conjunction with media gateways that use H.248-compliant media gateway control software to govern allocation of bandwidth to specific service demands. "The user is unaware of the mechanics of what's going on because the changes in network processes are performed on the fly based on the user's choice of a specific content option," Rashid says.
The service gateway at the CO provides the mechanisms by which various protocols in the IP protocol stack are activated across the fiber backbone into the CO or to the remote terminals to ensure appropriate service quality levels are available for, say, streamed content or communications applications, he notes. Intelligent agents keep tabs on the subscriber orders for billing purposes.
As some vendors focus on integration over the access link, others are providing the means for carriers to manage and provision any combination of Layer 2 and Layer 3 services from the core via new multiprotocol switches. The newest supplier to join the swelling ranks of vendors touting such switching platforms is startup Vivace Networks, which brought its first products to market at SUPERCOMM.
"Most of the companies in this space have tweaked their existing Layer 2 or Layer 3 switching platforms to turn them into multiprotocol platforms, but we've built ours from scratch with integration across both layers as the primary objective," says Mike Kazban, vice president of marketing and service at Vivace.
The company's core and edge switches support delivery of business class frame relay, ATM, Ethernet and IP services and can interconnect those protocols to ride over any given transport layer, Kazban says. "Our first deployment in a Tier 1 customer's network is to support an integrated ATM and frame relay service offering," he explains.
"We also go beyond class of service in IP mode by adding per-flow control in a way that's analogous to virtual circuit queuing in ATM," Kazban says. Among other attributes, the switches process frame relay at wire speeds all the way to 10gbps, which avoids the need to convert frame relay to ATM for switching when the transport input to the switch goes above DS3, he adds.
Vivace has its work cut out for it as entrenched competitors augment their multiservice capabilities. For example, Cisco Systems introduced its ATM Multiservice Portfolio, which includes the MGX 8950 core switch supporting 10gbps ATM interfaces and seamless interoperability between ATM and MPLS; a smaller switch for remote locations; and a route processor module, which acts as an integrated IP service gateway for the MGX switches. The combination allows carriers to upsell IP services like MPLS VPN and voice over IP to existing customers with traditional Layer 2 services from a single chassis.
At the same time, Cisco is taking steps to bring Layer 3 performance capabilities ever closer to Layer 2 levels via new software releases that provide significant resiliency gains in routing, MPLS and in link layers connecting routers with ATM, frame relay, Ethernet and other transport protocols. The upgrades, which began with a Cisco IOS release in June and will extend through the second half of the year, involve some protocols and techniques tied to new standards work at the Internet Engineering Task Force and others based on in-house refinements of existing processes.
One measure of improvement is in how Cisco routers perform BGP convergence, which is the updating of address records in routing configurations based on the Border Gateway Protocol. Optimizing BGP performance by better methods of communicating and processing updates in routers has produced dramatic gains, says Charles Goldberg, product line manager for Cisco's new Globally Resilient IP portfolio. "We tested to the volume of the largest BGP field that's in operation, involving 220,000 (IP address) prefixes, and found the routers converged 40 percent faster than before," Goldberg says.
Following release of this capability with the next IOS release, which was expected in June, Cisco will take convergence gains further by applying them to iBGP, which involves OSPF (Open Shortest Path First) and IS-IS (Intermediate System to Intermediate System) configurations. Cisco's new Incremental Shortest Path First Optimization technique refines the SPF routing algorithm so the convergence task is applied only to the part of the routing table that needs to be recalculated after a change in network topology. This can result in a 10 percent to 90 percent improvement in convergence rates on top of the gains from the initial BGP optimization release, Goldberg says.
The implications for Layer 3 performance are especially significant when it comes to what Goldberg refers to as the "crown jewel" in the new resiliency portfolio, which is the combination of techniques known as Nonstop Forwarding (NSF) and Stateful Switchover (SSO). "We're enabling the ability to switch over to the standby routing processor while forwarding packets and maintaining the state of Layer 2 connections that might be running on end user CPE or other neighbor devices," he says. Cisco anticipates the NSF/SSO features will be deployed at the edge of the service provider network.
NSF, involving protocols that work with BGP, IS-IS and OSPF configurations, allows Cisco routers with redundant routing processors on board to achieve zero packet loss during nonstop operations. This means neighboring routers don't have to update their routing tables, which adds to network stability, Goldberg notes. SSO assures that state information is transferred to the backup router processor during normal operation, thereby maintaining the Layer 2 connections after a switchover.
Clearly, it will take awhile for the carrier community to implement the many integration techniques now available to make data service feature provisioning a process analogous to voice feature provisioning. The beauty of all this technology is not only that it works over legacy networks, but it also allows incremental response to market demand.
Worldwide Multiservice WAN Switch Market
