Editor's Letter: Innovation and Legacies

It’s common – if not axiomatic – to link technology with change. That’s the widespread, popular and largely unthinking perception.

Among those who design, build and operate networks, however, that association runs into a cold, hard reality called legacy infrastructure. Examples abound.

Take Ethernet. It’s cheap and fast – and seems to get cheaper and faster by the day. Those are powerful drivers that have moved this technology into the metro and wide area network (MAN and WAN) space. But does that mean it will displace time division multiplexing (TDM)?

"Our story is that TDM services are not going away," said Fujitsu Market Development Manager Randy Eisenach earlier this year. "They’re going to be in the network for another 15 or 20 years."

As it happens, Fujitsu is banking on both TDM and Ethernet. The trick is a "universal" approach – in this case, a "universal switching fabric" – that can handle any type of traffic.

That word "universal" now frequently consorts with edge quadrature amplitude modulation (QAM) devices, as well. The pairing results from the elegant DOCSIS 3.0-related (and cheaper, faster) notion of an edge QAM modulator that can handle both data and video.

But this, too, runs into an installed base, a reality of which the engineers who specified the exacting, DOCSIS 3.0 downstream radio frequency interface (DRFI) were acutely aware.

"If we knew that all receivers would include modern DOCSIS 3.0 error correction, it would be a different spec," said Greg Taylor, Cisco Systems manager of broadband network engineering for the CMTS business unit. "We had to accommodate devices that were even pre-DOCSIS."

Accommoding to reality can be tough. "I would be surprised if anyone said that this (spec) is a simple thing to meet," said Taylor. But it’s within such constraints that practical innovation usually occurs.

Jonathan Tombes
Editor
[email protected]