Global Sources
EE Times-Asia
Stay in touch with EE Times Asia
EE Times-Asia > Networks

Packets devour time-domain nets

Posted: 01 May 2001 ?? ?Print Version ?Bookmark and Share

Keywords:tdm? ip? lan? wan? atm?

Every major provider of network infrastructure is developing equipment for packet-based networks, which are replacing traditional time-domain multiplexed (TDM) networks. This transition in network architectures is creating conflicting forces in the evolution of network equipment. The need to be compatible with legacy network infrastructure is competing with the need to provide greater bandwidth and greater user densities. To meet these needs, telecom equipment makers have had to fall back on a variety of architectures.

The existing Public Switched Telephone Network (pstn) is based on TDM switching, which uses a hierarchy of multiplexed channels, each carrying traffic at a rate of 56Kbps. The advantage of TDM, especially for voice traffic, is that each channel has a specific amount of bandwidth. The isochronous nature of the TDM network creates a deterministic amount of latency in the transport of the data. This simplifies voice transport algorithms like echo cancellation.

But the fixed nature of TDM is its biggest drawback for other types of data transmission; it means only so many users may use the network at a given time. When all the time slots are in use, the network is full. Moreover, the time slots are allocated when a connection is made, and are in use regardless of the actual amount of user traffic being carried.

Packet-based networks use routing and switching, and only transport data when data exists to transport. This makes more efficient use of the network infrastructure. Also, the growth of the Internet has led to much larger amounts of data being transported than voice. These two factors have prompted all major telecom providers to move toward packet-based networks.

There are a number of packet protocols, but the two that are the most dominant are Internet Protocol (IP) and asynchronous transfer mode (ATM). IP and ATM are significantly different in protocol and complexity. ATM uses small packets and supports a variety of quality-of-service (qos) routing mechanisms. This has the advantage of supporting different traffic types and optimizing network utilization. This is why much of the packet-based network backbone is ATM.

The disadvantage of ATM is this complexity. ATM-based equipment is often costly for a given amount of bandwidth. IP, by contrast, is a simpler protocol, reducing the cost of network deployment. IP networks are commonly used in LANs in homes and businesses. In these applications, bandwidth is substituted for network optimization.

Both ATM and IP are finding use in metropolitan-area networks. Many of these networks are ATM since the backbone is ATM, but as service providers also become owners of their own infrastructure, they have the benefit of being able to deploy IP networks if they choose. The mixture of IP and ATM packet networks and existing TDM networks create a need to interwork or bridge between these networks in many pieces of telecom equipment.

The real question is what transport to use in telecom platforms. TDM has significant capacity limitations. ATM offers great resource utilization and multiple classes of service, making it a good compromise between circuit switching and packet switching architectures-but the costs associated with ATM hardware and software is prohibitive for cost-sensitive applications. IP is much simpler, but QoS is not addressed in the basic layer, so additional processing is needed to deliver TDM-quality voice.

Another consideration is the network itself. If the application is IP-based and the network is IP-based, then an ATM transport layer would result in a lot of protocol conversion and a lot of cost. The same is true for using IP in an ATM application. There will continue to be competing network architectures for some time.

As a result, telecom vendors are faced with building many types of platforms and supporting many protocol layers. It is a significant disadvantage to develop and support all the different interfaces and protocols to convert between the different network layers. The costs and time-to-market penalties are significant. Also, service providers do not want to continuously replace installed equipment just to add new services.

The solution is a generic transport layer that can deal with many different networks and data formats. The ideal platform is one that allows a variety of interfaces and services to be integrated into the base platform. The base platform needs transport architecture with sufficient bandwidth for future growth and flexibility (protocol independence) to adapt to changes in market direction.

? Chuck Hill

Motorola Inc.

Article Comments - Packets devour time-domain nets
*? You can enter [0] more charecters.
*Verify code:


Visit Asia Webinars to learn about the latest in technology and get practical design tips.

Back to Top