Internet offload today, voice offload tomorrow

As data moves to the forefront, service providers need new alternatives for transport

Liberalization and deregulation are bringing new carriers with aggressive business models into the service provider market. As a result, there are tremendous pressures on tariffs that are forcing all carriers to reduce service costs. While competitive local exchange carriers are often in a position to build greenfield solutions that leverage the promise of IP, incumbent carriers are forced to defend themselves by drastically lowering their overall cost and shifting their investment for network expansion toward new, more efficient technologies.

These market forces are creating opportunities for carriers that integrate next generation switching technologies into their infrastructure to augment existing Class 5 switches or to create new networks using next generation technologies. By efficiently managing voice and Internet traffic, competitive and incumbent carriers can gain operational efficiencies and quickly develop new services.

Drastically changing traffic patterns

Internet traffic has been increasing dramatically and is projected to grow even more by 2005 (Figure 1). As a result, voice networks are becoming highly congested. This is particularly true of Class 5 switches, which were designed for voice traffic but are often being overloaded by data traffic. Adding new central office (CO) switches is expensive and, in many cases, impractical because of the lack of space at the CO to accommodate expansion.

Traditional CO switching technology was originally developed to support the reliability and availability requirements of predictable voice traffic flows. Carriers have been able to build on these platforms to deploy new voice features such as caller ID and call forwarding. However, with the rapid growth of data and Internet access traffic over the voice infrastructure, usage patterns have shifted and have negatively affected the existing economic models of switch use.

For example - although values vary by culture - voice call lengths in general used to average about 3.5 minutes per call. With the advent of remote access and dial-up traffic to the Internet, the average call now lasts 30 minutes or longer. Meanwhile, many of the value-added voice features are rarely being used. Voice switches that had been designed for predictable usage levels are being forced to accommodate longer data calls, and carriers must accommodate these shifts in caller behaviors.

As carriers worldwide face the need to expand to support data and Internet traffic, they are searching for other technologies that are more economical and can better help them meet future business challenges.

Wide open architectures

With data and voice being equally ubiquitous, IP technology is gaining importance in voice communication and new network architectures for public voice communication being developed. These architectures are based on a distributed functionality composed of media gateways and media gateway controllers.

Those different functions communicate through standardized protocols that mediate between infrastructures. While this trend from time division multiplexing-based systems to IP-based technology engages the public discussion, ATM is playing an increasingly important role, particularly in North America. It is not yet clear which technologies will become the most popular, so it is important to maintain the flexibility to accommodate diverse networking protocols.

This shift from circuit-switched to packet-switched networking creates the need for products that mediate between diverse networks. Mediation products must be able to leverage the cost-effectiveness of the data infrastructure as well as the reliability and quality of service of the legacy voice infrastructure. Therefore, key success factors include:

- Carrier-grade switching with five 9s reliability and the scalability to start with a capacity of a few hundred ports and potentially support up to hundreds of thousands ports

- Flexibility to support legacy and emerging protocols

- High-density platforms that efficiently use CO real estate, support more users in less space and make better use of CO air conditioning and power

- Lower costs that enable the cost-effective deployments of new services

Scenarios currently being discussed in the carrier space start with the immediate need for Internet offload. The dial-in traffic is groomed at the edge of the voice network and handed off to the ISP as early as possible - preferably before it hits the first transit exchange. New, alternative carriers in particular have the need to offload Internet traffic at the point of interconnection by deploying a high-density circuit switch in front of their transit exchanges (Figure 2).

Some carriers have found that up to 80% of the incoming traffic consists of modem calls that can be offloaded before hitting the precious CO switch. By deploying high-density service mediation switches, these carriers reserve legacy voice switches to carry the voice traffic for which they were designed.

The next step to optimize the network is saving the expensive primary rate ISDN on the switches and the remote access servers (RAS) of the ISP by integrating the RAS functionality into the service mediation switch. By terminating dial-up traffic at the service mediation switch, the IP packets go directly to the edge router for transmission over the Internet. This not only reduces the cost for the dial-up carrier, but it creates a more cost-effective business model for the ISP as well.

The complex and expensive remote access server functionality can be outsourced to the carrier while the ISP operates the functions related to delivering the service to the end user: authentication, authorization and accounting. The edge router can run the Radius protocol to the ISP's authentication, authorization and accounting server.

Voice gets upstaged

With voice about to become the minor share in overall traffic, telecommunications providers have to develop solutions for converging voice and data networks. In North America, ATM is the leading technology candidate, while in Europe and the Asia-Pacific, IP is more popular as a carrier convergence protocol.

The massive investments in circuit-switching infrastructure worldwide ensure that no operator in the foreseeable future will discard existing circuit-switched equipment. However, providers will most likely redirect incremental investments for capacity expansion and new services into next generation solutions based on IP or ATM.

Therefore, the next step is to convert circuit-switched voice into IP packets or ATM cells to provide for interconnection between the old circuit-switched networks and the new IP or ATM networks (Figure 3).

This will begin with low functionality such as capacity expansion of the backbone. Increasing voice traffic will be converted into IP packets or ATM cells and handed over to the carrier's IP network, which takes place in media gateways. When transit exchange capacity needs to be expanded, service providers will install programmable softswitches.

Media gateway controllers coordinate the media gateways so that the exact same functionality is provided as in the traditional transit exchanges. The motivation for this is to shift the call control from the proprietary CO equipment to commercial servers and split the call control function from the bearer handling traffic, resulting in cost reduction and faster deployment of new services.

In the same architecture, programmable softswitches will comprise local exchange functionality providing a ubiquitous platform for all converged services where voice and data applications are indistinguishably intermingled. Examples include click-to-dial and universal mailbox services.

The one-two punch

The converging voice and data market is being tackled from both sides, with IP infrastructure companies upgrading their products to add voice capabilities and traditional switch manufacturers migrating their CO equipment to IP-based, softswitch-like architectures.

Carriers therefore should understand the inherent bias of each approach. By selecting a service solution that is unbiased toward any particular technology, carriers can leapfrog their competitors and swiftly deploy new services while reducing infrastructure costs.