Accessing the Potential of IP Telephony
October 1, 2000
Posted: 10/2000
Accessing the Potential of IP Telephony
By Terrence Skemer
In the last two years, major carriers have announced significant spending on packet-based core networks. Verizon
Commu-nications’ (www.verizon.com) billion-dollar investment in metropolitan data networks and SBC Communication Inc.’s
(www.sbc.com) $6 billion packet-based buildout are two examples of this trend. These new network buildouts are not sport on the part of carriers. They are driven by end users–consumers with an explosive demand for enhanced, Internet-type services like point-and-click calling, online collaboration, video on demand and pay-per-use bandwidth. This demand is causing carriers to re-think the networks they build.
The industry’s attention is turning toward how to deliver these services across the access network, or the “last mile”–that part of the network that connects end users to their service provider’s first office. Carriers and vendors alike are working furiously to solve the challenge of how to bring those Internet-type services right to small offices and homes–with the level of QoS required to serve end users.
And the business case for doing so is clear. U.S. Bancorp Piper Jaffray
(www.piperjaffray.com) reported as early as February of 1999 that “widespread acceptance of IP telephony technology has moved from an ‘if’ market opportunity to a ‘when’ market opportunity.” Furthermore, International Data Corp.
(IDC, www.idc.com) forecasts that the size of the business market for IP telephony “will surpass the consumer market” by 2004.
Indeed, the delivery of voice and data services is probably the most rapidly growing opportunity for service providers, especially those targeting small and medium-sized business (SMB) markets.
Some argue that the key to enabling these Internet-type services is IP. But the issue is not over what protocol to deliver services. The issue is how to achieve the delivery of enhanced services without becoming bogged down by an infrastructure that will not enable them effectively.
So what’s the problem?
Before they can make IP telephony a business case that works, service providers must resolve the difficulties associated with delivering today’s broadband services across the access network.
Delivering enhanced services over any network presents serious challenges to local services delivery. Most significantly, quality suffers when voice or other real-time traffic, such as streaming video, is packetized.
Adding to the pain is the access network itself, where facilities are numerous and bandwidth is scarce. The many disparate pieces of network equipment that must be specially provisioned for every new service and customer both complicate and impede the effective delivery of broadband services across the access network.
In addition, because service providers have invested heavily in the circuit-switched networks that underlie all our cities and local services, they are left with a dilemma: how to deliver the enhanced services end users demand while still leveraging existing network investments.
Integration: The Investment Payoff
As service providers position to deliver differentiated services to the lucrative SMB markets, they are understandably seeking lower cost services and one piece of easy-to-manage equipment that can handle the transport of all a business’ required communications.
As the Yankee Group
(www.yankeegroup.com) explains, “An integrated network service should offer the end user reduced network complexity, easier ways to order and provision services, flexible billing options, and cost savings.”
At the same time, service providers are committed to integrating their existing infrastructures because integration allows them to provide the enhanced services that are in demand while reducing costs.
Other benefits of integration to service providers include the ability to:
Provision services more quickly;
Offer guaranteed throughput and assure service quality;
Penetrate new markets and increase revenues;
Strengthen the hold on existing customers and reduce customer churn; and
Realize cost savings, including: minimized operational costs through unified or reduced access network facilities, lowered cost of transmission, and lowered capital costs through the consolidation of traditionally disparate network elements.
In order to achieve this integration and have the ability to deliver enhanced services, service providers must strive for a “mass-customization” of services. Mass-customization refers to a business model in which a service provider leverages a core portfolio of services by tailoring them to fit the specific needs of differentiated vertical small and medium-sized business markets–such as financial, medical and hospitality industries (see
"Mass-Customization of Service" diagram).
Diagram: Mass-Customization of Service
The Access Network Factor
Mass-customization is blocked by the technical hurdles that prevent the delivery of advanced services over the access network.
The Yankee Group states that “unified networks, in most cases, are an unrealistic proposition today given the carriers’ investments in legacy networks [and] disparate networks obtained through acquisition.” Given this, carriers struggle with bandwidth-limited access infrastructures such as leased lines; multiple voice and data networks; and numerous disparate network elements to purchase, maintain and manage.
Access network solutions on the market today are collections of various network elements that perform specialized access functions (see
"Today’s Access Network Challenges" diagram). These include voice and data switches, IP routers, voice gateways and
IADs. When such a network is pushed to deliver high-speed broadband services, the multitude of elements results in ineffective traffic flow and impeded scalability. It also confuses the responsibility each element holds for functions such as traffic management and shaping, authorization and accounting, and QoS assurance.
Diagram: Today’s Access Network Challenges
In short, from business and technical perspectives, service providers’ current networks make goals like service differentiation and mass customization nearly impossible to achieve.
Most carriers have built their networks based on Class 5 CO switches because these provide reliable and intelligent voice service. Most also are planning to evolve to “switchless” next-generation networks. This situation presents two challenges.
First, how can a service provider offer local services based on Class 5 infrastructures in the near term for instant market penetration and revenue generation?
Second, how can the same service provider evolve the network and transfer customers and services to a next-generation network?
The Convergence Challenge
There are other challenges. As a report from FleetBoston Robertson Stephens
(www.rsco.com) explains, "Packetized voice still suffers from poor audio quality when compared to traditional POTS due to voice compression, packet loss and latency–that annoying pause you get on satellite long-distance calls.”
Because of this, transmitting voice over packet networks has not been very successful. In a data network, various sources of impairment affect quality. Traffic is fragmented to fit into packets or cells and reconstructed later; and packets are compressed and queued to optimize bandwidth and even dropped on a regular basis. The resulting jitter is compensated for with jitter buffers, which cause additional delay (see “Source of Packet Network Latency” diagram, page 96). These are acceptable treatments for non-real-time data, as delivery to the end user is not affected by average levels of jitter, latency and loss.
The trouble begins when packet networks treat voice frames like data–queuing, fragmenting and dropping time-sensitive voice frames in the same manner as they do with data packets. But voice and real-time traffic have a much lower tolerance for jitter and latency, which cause audio echo and overlap and choppy or fragmented video reception.
Finally, whereas data packet loss can be corrected through retransmission, this same method magnifies the problems that result from voice or real-time packet loss.
Contemplating an Enhanced Future
The solution to the challenges presented by the access network and the ability to achieve convergence is a services-aware architecture, such as intelligent edge switching. Intelligent edge switching connects the access network to the current Class 5 switch while also providing a migration path to interconnect with the Internet and next-generation networks with minimal service disruption and reinvestment.
A services-aware architecture operates transparently to the physical transport and data links. It effectively provides a point of convergence for core and access network infrastructures. This enables service providers to deliver integrated services over any infrastructure–such as T1, DSL, wireless or cable. At the same time, it reduces the management and provisioning burden inherent in traditional circuit-switched networks.
Using this approach, a service provider can choose any access infrastructure–or leverage an existing one.
Is this possible? Yes, if we leverage the inherent power of the access network.
A services-aware solution should use sophisticated queuing algorithms and packet conditioning to ensure that data transmission does not impact the delivery of real-time services such as voice and streaming video. In addition it should use simple and efficient packet labeling strategies to reduce protocol overhead, reduce IP address waste and achieve high throughputs.
Diagram: Sources of Packet Network Latency
Distribution of key network intelligence–such as authentication, authorization and accounting from the core to the network edge can minimize bottlenecks, effectively manage traffic, enable individualized service choices and self-provisioning, and increase the service provider’s ownership of the access network to provide end-to-end
QoS. Furthermore, application of voice compression techniques can be used to obtain superior-quality voice reproduction with minimum bandwidth consumption.
Diagram: Sources of Packet Network Latency
A services-aware network recognizes the traffic that crosses it as service specific. This allows for differentiated treatment of the traffic according to the QoS and priority associated with each service.
So although there is a revolution, it’s not an infrastructure revolution. It’s a services revolution. Implicit in the demand for “virtual” customized applications and self-provisioning is the end user’s desire for greater power over services.
In a complicated network environment where the infrastructure must be the focus of business, it is impossible to give the end user more control. A key to providing that control is to leverage the inherent and virtually limitless power of a services-aware access network. Solving the current challenges of the access network will open the floodgates to enhanced services.
Based on its flexibility and potential for reaching nearly every home and business in the world, the access network remains a powerful vehicle for service delivery.
Terrence Skemer is senior network architect at Sedona Networks Corp.
(www.sedonanetworks.com). He has spent more than 20 years in telecommunications networking and
engineering, and holds four patents in voice over packet and packet networking technologies. He can be reached at
[email protected].
Read more about:
AgentsAbout the Author
You May Also Like