Transport multiplexer design manipulates Sonet/SDH and ATM technologies
Transport multiplexer design manipulates Sonet/SDH and ATM technologies
A long-term add/drop multiplexer is designed to handle present and predicted high-data transport rates, as well as diverse processing technologies, within fiber-ring-based public communications networks
Future telephone and cable networks are expected to contain integrated, Synchronous Optical Network/Synchronous Digital Hierarchy (Sonet/SDH) and Asynchronous Transfer Mode (ATM) network elements. The
next-generation transport network design is projected to consist of a 155-megabit-per-second OC-3/STM-1 ATM access multiplexer at or near the customer location. In the fiber-optic loop ring, 622-Mbit/sec OC-12/STM-4 add/drop multiplexers are foreseen as the implemented equipment technologies. The future OC-12 transport add/drop multiplexer is likely to combine Synchronous Transfer Mode (STM) and ATM technologies in fiber-optic network arrangements.
The advantages of ATM technology stem from its cell-based structure, which allows ATM to carry voice, video and data communications in the same transport and switching equipment. Although ATM is anticipated to serve as the technology of choice for communications services, Sonet/SDH technology is expected to perform as the transport network.
A common misconception about ATM and Sonet/SDH is that they are competing technologies. Although some functions of both technologies contend with each other, future ATM networks are expected to use Sonet/SDH as the physical layer interface.
A key technology issue concerns network migration from the existing Sonet virtual tributary network to a Sonet/ATM network. Large amounts of Sonet/SDH equipment have already been deployed and should continue to be deployed in telephone company and cable-TV networks. Network providers want to position their platforms for the acceptance of ATM technology. However, they also need to maintain and enhance current levels of operating revenues.
To help determine future network architectures and migration paths, the workings of existing Sonet/SDH networks should be analyzed carefully. For example, a typical telephone network places OC-3 ring equipment in the loop feeder network. From the ring to the customer, a 1.544-Mbit/sec DS-1 line is usually extended either electrically, such as by a T1 carrier or by a high-bit-rate digital subscriber line, or optically by employing optical extension equipment, such as optical DS-2 at 6.132 Mbits/sec, OC-1 at 52 Mbits/sec or OC-3. At the customer location, a device such as a T1 multiplexer is generally used to combine traffic for network entry.
Future ATM networks are expected to markedly impact existing Sonet/SDH networks. ATM services drive the need for higher-capacity backbone networks to carry 150.336-Mbit/sec STS-3c traffic from customer locations. Once deployment of OC-3/STM-1 lines occurs at customer locations, the OC-3 loop network is expected to become obsolete, and OC-12 rates are expected to become the transport rate of choice for the loop/feeder fiber network.
In reality, though, even without ATM as a driver, future reduction in cost of optical interfaces and components will drive OC-3/STM-1 closer to the customer, eventually replacing the lower-rate optical extension equipment. Cost reduction is also expected to drive OC-12/STM-4 lines from the backbone network into loop applications that currently use OC-3/STM-1 equipment. ATM technology is anticipated to speed this migration process.
Moreover, the demand for high-capacity services, combined with the availability of DS-1 access from OC-12/STM-4 Sonet/SDH add/drop multiplexers, is making OC-12/STM-4 viable in the loop today. This viability becomes important because the OC-12/STM-4 network will be in place when customers require OC-3/STM-1-based ATM network access.
Due to these market drivers, the likely scenario for the future Sonet/ATM transport network is one at or near the customer location, and a Sonet/ATM access multiplexer located at the OC-3/STM-1 level (see Fig. 1). The OC-3 multiplexer feeds into an OC-12/STM-4 loop/feeder fiber-ring network.
One access method into the future network is foreseen as one in which ATM is carried over an OC-3/STM-1 line. The OC-3 access multiplexer takes the place of the existing optical extension multiplexer, whether Sonet or asynchronous. As applications emerge for larger bandwidth, such as for medical imaging, ATM customers are expected to demand OC-3 bandwidth from their locations. In these scenarios, the economical choice is the OC-3/STM-1 access multiplexer in a point-to-point arrangement.
In most applications, however, this arrangement does not prove cost-effective when the ring operating at an OC-3 rate is placed at or near a customer location when the customer needs OC-3-based service. Some scenarios do prove economical, however, when an OC-12/STM-4 ring is used.
Network providers recognize that some benefit arises from using ATM in existing applications--for example, in DS-1 circuit emulation. For these low-demand scenarios, the OC-3 access multiplexer is viable in fiber-ring architectures.
Bellcore has defined the service access multiplexer in its GR-2842 generic requirements as an ATM service access node typically deployed at the edge of a public ATM network. No other network element located close to the customer processes the ATM layer.
A private network equivalent to the service access multiplexer device performs the same functions, but is usually owned by the end customer. This device would be used instead of the public network service access multiplexer. In the transition, existing OC-3/STM-1 Sonet/SDH multiplexers can be modified to provide "access multiplexer" functionality and carry segmented STM/ATM traffic. However, in the long term, the access multiplexer is expected to be completely ATM over Sonet/SDH.
The basic function of a Sonet/ATM access multiplexer is to aggregate traffic from multiple broadband service access points and deliver this service traffic to an ATM switching system for service feature processing and switching (see Fig. 2). An access multiplexer is intended to perform as a low-cost broadband service access platform. By aggregating traffic to provide a highly filled interface to an ATM switch, an ATM access multiplexer allows better use of an ATM`s switch capacity, therefore reducing the switch cost per customer.
In the implementation of the Sonet/ATM access multiplexer, the SYN-155 (155-Mbit/sec synchronizer), the SOT-3 (STM-1/STS-3/STS-3c Sonet overhead terminator), and the cell delineation block provide the physical layer Sonet/SDH overhead capabilities for the ATM user network interface, network-to-network interface, or service access multiplexer interface. If required, all this functionality can be duplicated to provide survivability.
The cell delineation block-ATM interface supports key ATM network standards (not available in other commercial devices), including operations, administration and maintenance cell processing; universal test and operations physical interface for ATM; limited cell address screening; and interfaces to a variety of networks.
The multiplexing and switching of the access multiplexer is provided by the CellBus interface transceiver (Cubit) device. The CellBus is a 37-line parallel bus common to several network devices. Each Cubit device, together with a static random access memory chip, forms a complete section of an ATM switch. All the basic switching functions of cell routing, virtual path identifier/virtual connection identifier translation and cell buffering are contained in a Cubit device. Multiport ATM switching systems can be formed by connecting multiple Cubits over the CellBus.
In addition, the access multiplexer provides service interfaces. Two of the main near-term service applications are data and constant bit-rate services--for example, T1 and E1. For data applications, such as frame relay, a high-level data link controller device is combined with segmentation and reassembly devices to convert data into ATM cells for transport over the CellBus.
For ATM constant bit-rate services, existing DS-1/E1 lines can be converted using the ATM adaptation layer protocol using constant bit-rate ATM adaptation (Cobra) and framer devices. Cell-relay services can also be provided. For example, when an advanced DS-3 and STS-1 receiver/transmitter, DS-3 framer and cell delineation block devices are used, then DS-3 or STS-1 user network interfaces are provided.
Transport network migration
ATM transport over OC-3/STM-1 lines from the customer location requires a high-capacity backbone network. In the feeder/distribution network, the backbone network is likely to be based on OC-12/STM-4. The Sonet/SDH self-healing ring network is today`s technology choice for the transport network. Sonet/SDH technology has not only made the self-healing ring economically viable over traditional point-to-point architectures, but it has also dramatically improved network survivability.
Another misconception in the telecommunications industry is that ATM is inherently a point-to-point type of arrangement best suited for star architectures. In reality, because of its concentration capabilities, ATM is well suited for integration into self-healing rings. Consequently, the OC-12 ring is expected to be the technology of choice in locations now served by OC-3 rings.
For the near term, available STM rings offer ATM transport and survivability at minimal up-front cost. In the future, however, ATM-based techniques appear to be better solutions. They are expected to demonstrate survivability at both Sonet/SDH and ATM layers and allow low-cost implementation through integration and bandwidth utilization. u
Jonathan Morgan is Sonet/SDH product manager at Transwitch Corp. in Shelton, CT.