September 2004

Designing & Managing IP Networks for Voice (continued)

BY SAM SHIFFMAN

MPLS was a key component in our ability to build a redundant and self-healing network. In our plan, MPLS was one of the last things we did on the network to ensure there is not a single point of failure. Eliminating points of failure, takes your applications from three or four “9’s” of uptime to the voice telecom standard of “5 9’s” and true reliability.

While the network is the critical foundation which we built our house of services upon, softswitches make these services possible. Softswitches are moving us from an architecture where network control was embedded in proprietary hardware, toward a software-based architecture where hardware becomes more of a commodity decision point. A softswitch-based architecture is open and distributed in nature, allowing service and control to also be distributed. Softswitches represent low dollar investments that help minimize cost of ownership for legacy services and provide some of the capabilities necessary to introduced advanced services. In the early days of softswitches a room of computers was necessary to complete basic routing decisions, which was only slightly better from a space point of view than legacy “big iron” switches. Today a softswitch the size of your home computer can serve communication sessions of millions of users.

More recently, we’ve looked to utilize both softswitches and session border controllers, working in harmony, to enable voice and other real-time services over IP and accommodate cost-effective network expansion. Session border controllers deliver the advanced management functionality to overcome network security, signaling interworking, and multi-vendor interoperability issues that otherwise complicate interconnection of SIP- and H.323-based networks via IP. They enable a service provider to more easily interconnect with carrier partners via VoIP peering.

Traditionally, service providers have used expensive TDM ports on media gateways to interface with other VoIP service providers. In TDM peering, DSPs on media gateways linked back-to-back convert voice traffic from VoIP to TDM and back to VoIP again at the network edge. In VoIP peering, however, the service provider converts voice traffic to TDM only when it originates or terminates calls on the PSTN or other TDM-based networks. Also known within the industry as “packet peering,” this capability can reduce the capital expenditures required for service-provider interconnects by 50 percent, depending on capacity, because DSP resources are no longer needed.

Furthermore, with session border controllers at the network edge, a service provider can harvest TDM switch ports used for VoIP-to-VoIP traffic and re-deploy them for TDM traffic. It also enables the provider to better utilize switch and media gateway ports used for TDM-to-VoIP traffic. So, in addition to enjoying CAPEX savings for new VoIP traffic, deployment of session controllers enable the provider to delay purchase of additional ports on TDM switches and associated media gateways.

Session border controllers simplify service provider interconnects; eliminate signaling problems such as message formatting, timing, and industry interpretation, and improve network address translation (NAT) and port address translation (PAT) abilities. Session border controllers also enable our network to route and manage authorized traffic to optimum peering points. The result is a more scalable network, delivering improved call-completion rates.

Today, session border controllers enable cost-efficient VoIP peering and delivering advanced traffic-management capabilities end to end across the network. This synergistic approach — leveraging both softswitch and session-control technologies — is eliminating traditional scalability and manageability limitations and spurring accelerated adoption of VoIP and other advanced services. Border elements of the future will also have to take on new critical functions that not only allow advanced IP communications networks to interconnect but also share network information with one another.

Are we at the end state? Absolutely not! Our starting goal was to build a foundation that was basically bridge between the old (PSTN) and the new (IP). We’re at point now where we have built the core foundation of a network that is extremely stable; IP networks now work as good as or better than PSTN. From that infrastructure, it is now possible to build enhanced voice applications that would never be possible in the circuit-switched world. Going forward, the technology will continue to evolve to push intelligent functionality further and further to the edge of the network.

Page 2 of 2 [Back to Page 1]

Sam Shiffman is executive vice president of PointOne of Austin, TX. For more information, please visit the company online at www.pointone.com.

If you are interested in purchasing reprints of this article (in either print or HTML format), please visit Reprint Management Services online at www.reprintbuyer.com or contact a representative via e-mail at [email protected] or by phone at 800-290-5460.

[ Return To The September 2004 Of Contents ]