Previous Table of Contents Next


NHRP

Another technology designed to allow multimedia to be carried over IP is the NHRP being developed by the IETF. NHRP aims to eliminate or bypass some or all of the routers used by Classic IP over ATM by directly connecting to the ATM fabric. Disjoint IP subnets are viewed as one logical NBMA. Each NBMA has a server that resolves IP addresses to NBMA addresses using dynamic address-learning mechanisms. Once addresses are resolved, a node then connects to the destination node with the required QOS parameters.

Although NBMA servers reduce the number of hops, they may increase the response time by a round-trip time (a critical factor for some applications). NBMA servers add network management complexity and introduce additional points of failure in the system.

Switched Multimegabit Data Service (SMDS)

SMDS is a connectionless service to interconnect LANs that enables access at T1 and T3 speeds. It can slice up to 9K bytes of data into fixed 53-byte cells (same as ATM) that are then switched through the network. Because it supports up to 9K bytes, it allows entire LAN frames to be encapsulated and provides reliable interconnections between LANs. The advantages of SMDS include scalability, bandwidth on demand, connectionless service, and multicasting support for some protocols.

Suitability for Multimedia Traffic. Because SMDS is packet-based, the delivered data stream can experience delays. This delay is not large and is less than 30 microseconds. In addition, SMDS does not provide for any synchronization of the data. The application is responsible for multiplexing audio and video and synchronizing the data at either end. The choice of 53-byte cells was made to provide a migratory path to ATM technology, rendering SMDS an interim technology.

Fiber Channel

Fiber channel is a new technology that integrates the channel technology of mainframes and networking fiber technology. Fiber channel provides a single standard for network storage and data transfer and moves device interconnection and switching to a fabric. Each network node is responsible only for the single point-point connection between itself and the fabric. The fabric is responsible for routing between nodes, error detection and correction, and flow control. It supports distances of up to 10K meters at speeds of 266M bps to 4G bps.

Suitability for Multimedia Traffic. Applications that are compatible with ATM run on fiber channel, which offers an even shorter delay than ATM. Because fiber channel integrates both storage devices and networking, it is ideal for use on servers such as video servers. To bring the technology to the desktop is costly, however, running roughly $2,000.

Because of its limit of 10K meters, fiber channel is really more suited to LANs or MAN. Unlike ATM technology, the entire standard for fiber channel has been formalized and interoperability is not an issue. Exhibit 5 compares the ATM and fiber channel technologies.

Exhibit 5. Comparison of ATM and Fiber Channel
ATM Fiber Channel

Congestion Handling Discards Not an option
Multivendor and Multiprotocol Struggles Handles easily

Suitability of WAN-based ATM

As previously discussed, ATM has been designed from the ground up to carry multimedia traffic on the WAN.

Although ATM provides the capability necessary for delivering multimedia traffic, much of the backbone or desktop connections have not been implemented. Making a desktop unit ATM-capable costs around $1,000. ATM on the WAN costs an average of $100,000/month when traffic exceeds 1.54M bps. Another issue that must be resolved is the interoperability of the switches from different vendors. Standards (in addition to PNNI) are still evolving to provide this and other capabilities. PNNI automatically disseminates network topology and resource information to all switches in the network and enables QOS-sensitive cell switching.

Despite these issues, ATM’s support for real-time and data applications, quality of service, and multicasting, and its low delay and jitter make it the perfect match for multimedia communications. ATM’s ability to be the transmission system for multiservice networks — those designed to carry voice, data, and video — provides profound advantages to a business by reducing the costs for equipment and support.12 Because ATM also allows great flexibility and every other multimedia technology provides an access path to it, ATM is the logical long-term choice networking solution.

CONCLUSION

Each available multimedia technology should be considered one of the many tools in the manager’s kit. Each technology can be mixed and matched to suit the enterprise’s networking needs (e.g., bandwidth, priority, and availability) and budget. Cooperation between network architectures and application architectures is important for predetermining traffic patterns and proactively managing bottlenecks.13

Almost all multimedia networking technologies provide for interconnection and incremental deployment. Because early implementations of any technology may mean interoperability issues have yet to be resolved, cautious integration is prudent.

Developing a strategic plan for multimedia networking that meets the goals and objectives of all users should encompass both computers and communications needs. Because voice and video communications will in all likelihood be a part of the network plan, users of these applications should be included in the planning and implementation team. To reduce the disruption that accompanies a technology transition, careful attention should be paid to migration planning.

Resource planning and project planning software help estimate the network resources needed to deliver a multimedia solution. Implementation should also take place in stages. For example, a workgroup solution may first be installed and pilot-tested before introducing the backbone technology. Network management is another area that mandates careful planning. It should encompass planning for configuration management (i.e., network element administration to provide services), fault management (i.e., detecting and isolating faults in network elements), and performance management (i.e., monitoring performance and managing traffic).

References

1.  N. J. Muller, “Multimedia over the Network,” Byte, March 1996, pp. 73-80.
2.  N. J. Muller, pp. 73-80.
3.  H. J. Stuttgen, “Network Evolution and Multimedia Communication,” IEEE Multimedia, Fall 1995, pp. 42-59.
4.  Stuttgen, pp. 42-59.
5.  D. Ruiu, “Testing ATM Systems,” IEEE Spectrum, June 1994, pp. 25-27.
6.  J. N. Rodriguez, “ATM: The New Communication Era,” IEEE Computer, May 1995, pp. 11-14.
7.  N. Kavak, “Data Communication in ATM Networks,” IEEE Network, May/June 1995, pp. 28-37; R. Jeffries, “ATM LAN Emulation: The Inside Story,” Data Communications, Sept.21, 1994, pp. 95-100.
8.  I. J. Hines, ATM: The Key To High Speed Broadband Networking, New York: MT Books, 1996; W. Goralski, Introduction to ATM Networking, New York: McGraw-Hill, 1995.
9.  N. Kavak, “Data Communication in ATM Networks,” IEEE Network, May/June 1995, pp. 28-37.
10.  R. Jeffries, “ATM LAN Emulation: The Inside Story,” Data Communications, Sept. 21, 1994, pp. 95-100; W. Goralski, Introduction to ATM Networking, New York: McGraw-Hill, 1995.
11.  K. Nahrstedt, and R. Steinmetz, “Resource Management in Networked Multimedia Systems,” IEEE Computer, May 1995, pp. 52-63.
12.  R. Madge, “Technology’s Real-World Edge,” Information Week, Sept. 16, 1996, p. 178.
13.  E. Horwitt, “IP over ATM,” Network World, April 15, 1996, pp. 40-42, 45.


Previous Table of Contents Next

Copyright © CRC Press LLC