SAN MATEO, Calif. — Conexant Systems Inc. and Metalink Ltd. are spearheading an effort to ram a common format for symmetric digital subscriber lines (SDSL) through the often-tedious standards-setting process, in an attempt to bypass what they consider a blockade by traditional telephone companies. But though the companies state their technology is already a de facto standard and should serve as the basis for an official standard, unresolved technical issues add an element of uncertainty to the format's adoption by the American National Standards Institute.
"We're going to attempt a blitz standard," said Danny Gur, vice president for business development at Metalink. "Having a standard in place will accelerate SDSL deployment in the United States."
SDSL technology offers digital bandwidth of up to 2.3 Mbits/second over a single twisted-pair copper phone line, and is aimed at the corporate and SoHo markets that require high upstream and downstream traffic rates. Gur said that Metalink's chips are compatible with those from Conexant, and that the common format dominates the installed base for SDSL systems, though Conexant commands the lion's share of the market.
Metalink's customers include several large OEMs that provide SDSL systems, and those boxes have in turn been purchased by numerous companies that want to offer DSL services to end users, Gur said. But there is a problem with this plan, Gur said, because telephone companies control access to the copper wires that connect homes and businesses with the central office sites where DSL systems reside.
No current standard exists for SDSL data transmission, and the phone companies so far have balked at deploying the expensive systems without one, Gur said, mindful that the emergence of a different format as the official standard at a later date would make purchases of the first units expensive mistakes.
Artful dodge?
But Gur called this excuse a dodge that allows the phone companies to delay the implementation of SDSL and extends the lifespan of T1 and E1 technologies. T1 and E1 are the most common high-bandwidth technologies currently used in the corporate market, and they are an important revenue stream for telephone carriers.
"Everybody is already using our technology," Gur said. "It is a standard in everything but name. Our technology allows full T1-class service but is much less expensive. T1 service is a bad deal for everybody, but the phone companies want to slowly deploy DSL service and keep their T1 business intact as long as possible."
The plan is to use the upcoming ANSI meeting in Hawaii as a forum where the Conexant and Metalink, along with OEMs and DSL service providers — all of whom are basing their systems around the Conexant/Metalink architecture, Gur said — can approve the format as the ANSI standard for SDSL. While such standards-creating processes are often long and complicated, he said the group is not likely to encounter any opposition at the meeting and the standard could be approved before the end of this month.
Force feeding
If the standard is set, it could eliminate the phone companies' argument against SDSL deployment and force them to rapidly implement the technology. "We are going to ram it down their throats," Gur said.
However, the problem may be more complicated than simply creating working chips and convincing OEMs to buy them. "Even if there is a common format at both ends of the line, there can still be interference on the lines and that is a very serious issue," said Ken Krechmer, technical editor for Communications Standards Review (Palo Alto, Calif.).
Krechmer said that twisted-pair cables are generally bundled into groups of 25 lines, and that the various types of digital service, including ISDN, T1, E1 and the alphabet soup of numerous DSL formats can easily cause interference within these bundles simply by placing the lines next to each other. Each type of technology has its own power spectral density (PSD), and the phone companies have created seven PSD classes in order to study and prevent line interference. And any type of new standard will have to solve both the architectural issues found in the systems and the PSD interference problems found in the lines that connect the boxes, Krechmer said.
"For the chip companies to say that they are home free in the standards process is a bit naive, because creating a standard is a very complicated process," said Krechmer. "While the phone companies have often used marketing reasons as the basis for their decisions on whether to deploy a technology, in this case there are some very significant technical issues to be resolved."
