New Breed IAD Stands on the Edge

Pushing network intelligence into the customer premises

Mark Veil

To deliver converged services and graded SLAs for data applications, VPNs and Internet access cost effectively, carriers must stretch network intelligence from the edge into the customer premises. Fortunately a new breed of intelligent IADs (integrated access devices) are leveraging established and emerging traffic management and engineering concepts to enable network operators to deliver traditional, as well as packet-based, voice and tiered-data services profitably over a single access network. Using standards-based technology, these IADs help create new revenue opportunities and reduce operational costs.

Today, the public network is the core infrastructure of virtually every business, large and small. E-business is no longer a concept or catch phrase, it is a way of life. As a result, business requirements are fueling evolution and innovation in the public network for new services such as data and voice VPNs, streaming video, and voice-over-packet applications. To meet these demands, legacy voice and data networks are headed for convergence onto a common, ubiquitous, multipurpose IP-based platform.

Traditional IP networks, however, operate on a connectionless, best-effort basis, with all packets subject to equal treatment as they are routed individually throughout the network, hop-by-hop to their ultimate destination. This best-effort model of fairness translates to relative unfairness for traffic that is more sensitive to network impairments and doesn't align well with business plans that call for delivery of a rich portfolio of differentiated services and applications.

Consequently, delivering revenue-generating applications over converged, IP-based infrastructures requires a new breed of access network: one engineered to deliver carrier-class service but optimized to associate traffic streams with their applications and process each according to its predefined SLA. This new IP network must provide the same--or better--service quality than existing infrastructures.

Demarking a New Edge for QoS

To ensure that each service receives the appropriate QoS treatment and meets SLA obligations, IADs must manage, monitor and control network traffic at the service level, i.e., provide advanced traffic management and engineering services.

Specifically, IADs must ensure that QoS objectives are satisfied for new and existing traffic flows and protect against congestion and degradation of network performance. They must monitor and control the latency, jitter, average and peak packet rate, and packet loss ratios to ensure that availability and performance is within acceptable or contracted service bounds, and that premium or priority services are given preferential treatment. To achieve this, the IAD must provide facilities for traffic classification, admission control, traffic shaping and rate control.

Packet classifiers within the IAD must map packets requiring the same or similar QoS treatment to specific outbound queues. These traffic classifications are based typically on the contents of the packet header, such as the Layer 2 and Layer 3 source/destination address. In practice, however, classifications may be derived from (and applied to) a virtually unlimited range, combination and granularity of packet attributes, including physical ingress port/interface, application protocol type or IPv4 ToS (type of service) and IPv6 CoS (class of service) markings.

Admission control services within the IAD must ensure that the requested traffic profile and QoS levels be met concerning current network state, resource availability or other policy-based considerations prior to admitting the traffic flow. In addition, a variety of traffic-shaping and conditioning mechanisms must be employed to monitor and maintain compliance with traffic profiles or contracts. Finally, metering services must monitor and measure traffic against its profile and pass packets along to the appropriate policing mechanisms: the queuing and dropping services.

Connection Flows vs. Packet Hop

Once the IAD has classified and groomed the service flows appropriately, traffic engineering services must be applied to aggregate and map them efficiently onto the existing network topology to control network behavior, optimize network resources and maximize traffic delivery performance.

In heterogeneous public networks, MPLS (multiprotocol label switching) represents the best alternative for enabling IADs to perform traffic engineering and manage QoS (see Figure 1). Although originally intended to enhance routing performance, continued improvements in that area have shifted the application focus of MPLS to its inherent capabilities for delivering efficient and scalable traffic engineering and QoS in IP-based networks. Since MPLS operates at Layer 2.5, is protocol-agnostic, and separates forwarding and control functions cleanly, it supplies the intelligence required to associate a traffic stream with its service and process it according to the specified traffic contract.

MPLS gives IADs the ability to associate and allocate any type of traffic with a particular FEC (forwarding equivalency class). Each FEC represents an aggregation of traffic that will be treated in the same manner as it traverses the network. These FECs are mapped to LSPs (label switched paths) that have been engineered to support specific SLAs (e.g., guaranteed bandwidth, low latency). LSPs behave like the more familiar ATM virtual circuit and frame relay DLCI (data link connection identifier), but with greater efficiency.

Managing at the Service Level

By using IADs to extend MPLS to the customer premises, carriers can apply virtual routing to the local loop and enable a connectionless IP infrastructure to support connection-oriented services. By classifying, mapping and aggregating ingress traffic into service and/or application level virtual connections or LSPs, providers can manage at the service level.

In the TDM or ATM/PVC world, available access capacity is allocated to voice and data services in fixed apportionments according to their unique application requirements. Consequently, any unused capacity cannot be used by any other service and results in idle, wasted resources. For example, pauses in voice conversations cannot be filled with high-priority data traffic, and lulls in CIR (committed information rate) data transmissions cannot be exploited by best-effort traffic.

In sharp contrast to TDM and ATM's typical PVC implementations, this new virtual routing model allows carriers to reclaim fully all unused portions of access network capacity. Moreover, rather than nailing-up statically provisioned pipes, new IP/MPLS packet-based IADs allow providers to allocate access network bandwidth dynamically and generate new revenue opportunities from what would be wasted resources (see Figure 2).

By supporting both a physical and logical distribution of network intelligence, IADs create access networks that are feature-location agnostic. This virtualization of the access network enables carriers to deliver extremely scalable, efficient and secure private voice and data networks and transparently drive voice and unified communication features directly to the customer's doorstep. Intelligent IADs reduce the complexity and operational costs associated with operating multiple networks for each service and provide a single IP infrastructure that creates opportunities for bundling products, single billing, and developing new services that leverage voice, data and IP functionality.

Top of Page