Persevering Asynchronous Market Offers Realistic Perspective on Chaotic Industry
By: Mark Lutkowitz, <%=company%>
When Adtran (Huntsville, AL) announced the introduction of an M1-3 multiplexer on September 14, 1999, the press reported the story in a matter-of-fact fashion, with no fanfare. In an industry that has made synchronous optical network (SONET) technology synonymous with add/drop multiplexing, it did not seem to matter much that a supplier perceived a need for a new asynchronous system.
The Adtran product is definitely not consistent with the rhetoric of next-generation networking. But every now and then, prudent industry observers should step away from the clamor of emerging markets, at least half of which will amount to commercial vaporware for several years, and take a look at mature technology. Observing the life cycle of very mature markets offers a more realistic perspective on the rate of change in the optical telecommunications market than the overwhelming clamor for next-generation attention.
The market for M1-3
M1-3 devices that take 28 DS-1s and multiplex them into a DS-3 have not been getting the attention that they deserve for at least the last four years. Despite the perception that SONET terminals have totally consumed the M1-3 capability over this time period with ubiquitous deployment, there has been a healthy market for M1-3s. Significant price erosion will contain the M1-3 market to slight growth over the next few years in terms of dollars (see Exhibits 1 and 2). Shipments are made mainly to the U.S., with a small amount going to Canada and Asia.
In addition to Adtran, a company that has been very instrumental in driving M1-3 prices down, active suppliers include Carrier Access Corp. (CAC, Boulder, CO), NEC Eluminant Technologies (Herndon, VA), and Telco Systems (Norwood, MA).
Adtran was "pleasantly stunned" by the very fast level of acceptance of its new product. Eluminant claims it is still seeing "near record sales" for M1-3. CAC, a favorite of the financial markets, owes much of its success to its M1-3 business. After issuing a release of a new M1-3 product in the spring of 1998, Telco had shipped 7,000 units by September 1999, and is probably up to about 20,000 units by now. It also appears that one or more of the traditional transmission players, including Lucent, still sell older types of M1-3 products.
M1-3 trends
The largest growth segment of the M1-3 market has been sales to competitive local exchange carriers (CLECs). Some CLECs that have lacked sufficient capital to build SONET infrastructure have viewed these asynchronous products as a cheap means of transport. After these service providers interface with Interexchange Carriers (IXCs) or Incumbent Local Exchange Carriers (ILECs) at DS-3, they run services at DS-1 out to their customers.
Some of the CLECs wielding M1-3 transmission may just have a small customer base or a low volume of provisioning. Others may be prone to put up dedicated circuits and not do a lot of switching, which obviates the need for intelligent SONET cross-connect functionality. One port on a cross-connect can cost as much as the average selling price of an M1-3. Furthermore, the restricted space requirements of the CLECs, especially those in central office co-locations, makes the one-rack unit size of the M1-3 an attractive option compared to the large digital cross-connect systems (DCSs).
Old habits
Despite corporate strategies and internal policies among carriers to move away from older technology, old habits are hard to break. To replace analog switches, some ILECs still use M1-3s as an interface between the new digital switches and transmission gear. At least one of the Bell companies had shifted to M1-3s because of the capacity restrictions with tying multiple wideband (3/1) DCSs together at the same location. At other times, capital constraints dictated spending only when there was customer demand, so if service was needed within a week, the M1-3 muxes were more desirable than the much longer deployment lead time required by SONET gear.
Even in good times, given the much greater quickness of deployment, M1-3s have been the pragmatic choice of engineers as a short term solution to shore up their networks. A particular IXC might prefer to take advantage of the network flexibility and survivability offered by a SONET DCS, but the carrier sometimes would need to grow its capacity faster than it could install these cross-connects. Although the IXC may incur a much greater expense in upgrading to SONET in the long run, the near-term revenue from large customers could justify the initial M1-3 deployment.
New requirements
In the short term, M1-3 suppliers will continue to invest in development of products that are even smaller and less expensive, as semiconductors with higher levels of integration become available. The suppliers will also place more emphasis on value-added features.
Increasing demand for STS-1 feeds will make M13s more complementary with SONET gear. These synchronous interfaces will result in more efficient and flexible hand-offs to add/drop multiplexers, as well as offer increased SONET visibility. E1 drop capability on M1-3s for international gateway applications also will become more common. Other possible enhancements on M1-3s include improved remote management as well as performance monitoring, greater protection, additional types of interfaces (such as links to LANs as well as point-to-point fiber interfaces) and cross-connect functionality.
Proven technology perseveres
Although the current impact of the M1-3 market on the SONET DCS manufacturers is relatively small and practically unnoticeable, the key point is that the M1-3 market's very existence demonstrates the very slow rate of obsolescence in staple telecommunications products. Observing a real and long-established market such as that for M1-3 makes the fashionable predictions of the imminent demise of SONET that much more absurd. It is highly probable that a SONET equipment supplier, such as the venerable cross-connect house Tellabs (Lisle, IL), will continue to make lots of money selling its TITAN product for a very long time.
Many years from now, history is liable to repeat itself. Next-generation network proponents will suggest that wide-scale deployment of a very sophisticated, state-of-the-art technology will replace the dense wavelength division multiplexing (DWDM) systems that scale capacity in networks today. This distorted outlook will not take into account that there will still be a strong market for what will, by then, be a very old and extremely cheap technology–SONET–that did not get totally displaced by wavelength networking.