T-1: Dodo Bird or Phoenix?

With all the talk about Internet protocol (IP) and multimedia, you’d think the days of T-1 are gone. Vyyo, Narad Networks and Scientific-Atlanta are betting development and marketing dollars to the contrary and are trying to convince cable operators that money can still be made from this vintage technology. Recently, I asked these vendors why operators should invest in solutions to support T-1 and how those solutions play into the emergence of multimedia, IP Multimedia Subsystem (IMS) and seamless wireless-landline integration. T-1 review T-1, as you may recall, is the granddad of digital multiplexing. It is capable of a whopping data throughput of 1.54 Mbps and carries 24 digitized voice channels. As digital communications evolved, it became necessary to fit enhanced signaling and operations support into a T-1 frame, in addition to voice information and markers that determine where frames begin. The technology went through three major iterations, which first defined the basic frame structure, then a 12-frame superframe, followed by the 24-frame extended superframe. Knowing where to find all this information depends upon dedicating regular time slot intervals to each of the 24 voice channels and maintaining frame timing across networks within strict tolerances; hence, the name for the synchronous technology behind T-1: time division multiplexing (TDM). Unlike TDM, IP packets are typically sent as needed, rather than at regular intervals. This asynchronous mode of transmission is a more efficient use of bandwidth. Also, packet structure is inherently more flexible than T-1 frame structure, since packet length can vary, and headers define the characteristics of the information being transported, including type of information, signaling, error correction, and addresses for sender and receiver. Why T-1? So why would any one vendor, much less three, market new products to deliver old T-1 technology? Brent Levetan, Vyyo’s VP Business Development, provides an explanation: "T-1 is a broadly installed technology and will be for a very long time. Our national infrastructure is based upon SONET (synchronous optical network), and T-1 is part of the SONET hierarchy. Hundreds of billions of dollars have been spent on this infrastructure. Two major applications that cable can serve well are enterprise PBX (private branch exchange) and cellular site backhaul." Levetan went on to note that despite advances in IP as a transport protocol, the majority of enterprise PBXs represent too much capital commitment to be easily replaced and still require T-1 interfaces. He noted that although cellular backhaul is not as large a base as enterprise network interconnection, the growth rate of cellular is still in the neighborhood of 8 percent per year, and T-1 is the prevalent technology for linking cellular tower sites to base stations. Vendor solutions If the existing base of users is sufficient motivation to serve these markets, the challenge is to provide a solution that is compatible with DOCSIS, which is an asynchronous technology, and T-1, which is synchronous. Vyyo, Narad, and Scientific-Atlanta each came up with different solutions. The Vyyo solution is designed specifically for T-1 service using DOCSIS chipsets, while the Narad and S-A offerings are nonDOCSIS symmetrical Ethernet services that can transport T-1 over a DOCSIS-based cable plant. Vyyo’s solution consists of DOCSIS-compatible V300 T-1 modems at customers’ premises, connected over an HFC network to their headend-based XMTS T1 Access Concentrator. Both the modems and the access concentrator use Broadcom DOCSIS chipsets, but also proprietary technology to maintain T-1 timeslots and clocking. The XMTS interacts with the modems and a backbone transport network similarly to the way a cable modem termination system (CMTS) works in a DOCSIS network. Narad’s offering is an Ethernet switched solution that consists of Optical Network Distribution Switches (ONDSs), Narad Diplex Filters (NDFs), Narad Access Switches (NASs), and Broadband Interface Units (BIUs). The BIUs are customer premises-located 100BaseT Ethernet modems that contain a pseudowire interface that converts T-1 to IP for transmission over an Ethernet network. NASs are Ethernet switches that bridge existing amplifiers and regenerate the Ethernet data. The ONDS is a switch located at a fiber node and is where the T-1 is switched to a fiber distribution network. Narad uses an upstream carrier in the 908 to 971 MHz band and a downstream carrier in the 1.02 to 1.08 GHz band to provide symmetrical data service and depends upon the NDF to separate Ethernet from the standard cable spectrum at the ONDS. Like the Narad product, S-A’s BroadLAN Transport System is Ethernet-based and requires a T-1 to Ethernet interface from a partner company. Unlike Narad, it operates within standard upstream and downstream allocations. The hardware consists of a headend-based T-1200 or T-2400 Transport Modem Termination System and the customer premises C-1100 or C-1200 Client Transport Modem. It uses a 6 MHz channel in the 88-864 MHz downstream spectrum and another 6 MHz channel in the 5-42 MHz upstream. In the reverse path, subscriber information is multiplexed on up to 14 narrowband tones. The multiple tones provide immunity from "jammer" frequencies within the 6 MHz band. Future directions Although the rationale differs, each vendor maintains its version of T-1 connectivity is fully compatible with emerging multimedia, PacketCable and IMS technologies. Vyyo’s Levetan commented that, since their solution is DOCSIS-based, they are committed to updating the product as DOCSIS and PacketCable evolve. Chuck Kaplan, Narad Networks COO, noted that his company’s solution is independent of DOCSIS, and therefore compatibility was not an issue. Mark Palazzio, S-A’s VP Access Networks, echoes this sentiment for BroadLan, but also concedes that alternative TDM over DOCSIS solutions might emerge as DOCSIS 3.0 comes into the market with channel bonding and other bandwidth enhancements that could facilitate symmetrical services over DOCSIS. Like many commercial services applications, it’s still early in the lifecycle for T-1 connectivity and difficult to predict how much of this market will actually go to cable. These solutions were all shown at Expo in mid-2005, and as of this writing, the vendors are still saying that "several" operators are conducting trials. Vyyo was able to publicly cite one implementation each of cellular backhaul (Hargrade, in North Carolina) and enterprise connectivity (D and E Communications). Narad indicated Cablevision, Delta Cable and CAI Westland in Europe use its product, although not necessarily for T-1 connectivity. S-A preferred to let its customers announce any use of its product. Also, despite Levetan’s comments about the installed base of PBXs, claims of large potential in the enterprise market must be tempered by trends toward much improved scalability and pricing of IP PBX solutions. Whaleback Systems, for example, made a splash at Emerging Technologies with a buy-as-you-grow, flat $45 price per line for both equipment and usage that starts as low as three lines and can grow to 5,000 lines. Hedging bets As a closing note, it’s interesting to observe that although the vendors are publicly optimistic about the prospects for T-1, they are still hedging their bets. Both the Narad and S-A products are not restricted to this application, and are, in fact, fully functional HFC-compatible Ethernet switches that can also provide symmetrical data solutions over cable plant. Even Vyyo’s system, although marketed primarily as T-1 connectivity, has a 100BaseT port on its V300 modems. Justin J. Junkus is president of KnowledgeLink and telephony editor for Communications Technology. Reach him at [email protected].