Avoiding the proprietary path
A top technical mind recently expressed to me a worry that the cable industry has grown so large and cumbersome that it, if it's not careful, could find itself getting bogged down to the point that it will be incapable of ushering in innovation.
Although 2.0b, as its critics maintain, could slow the progress of DOCSIS 3.0, it's abundantly clear that there are competitive needs for cable to enlist downstream bonding now, rather than when it's too late. But I think it's also smart to insert CableLabs into this process, even if it's in the role of providing the test plans and ensuring interoperability with past and future versions of DOCSIS. This will also give operators the assurance that they won't end up with a bunch of proprietary equipment and stranded capital once DOCSIS 3.0 takes shape.
Having that involvement at such a high level should also provide some promise that 2.0b won't become the non-starter that DOCSIS 1.0 "Plus" was in the early part of the decade. Although never officially sanctioned by CableLabs, "Plus" used a proprietary method to inject quality of service (QoS) in DOCSIS 1.0 plant well before DOCSIS 1.1 equipment, delayed as it was, arrived. DOCSIS 1.0 "Plus" was used sparingly for early cable VoIP deployments and trials, and hardly qualified as even a stop-gap.
But 2.0b also applies some risk to the overall vendor equation. At the chip level, Broadcom Corp. appears to be in an advantageous position with 2.0b, as it surely has by this time introduced a new cable modem chipset designed to bond three downstream DOCSIS 2.0 channels. Texas Instruments, which is sure to play a key role with DOCSIS 3.0, will not play in the 2.0b pool (mostly due to instruction from operators that 2.0b is just for the short-term, and that TI should continue to focus on 3.0).
But 2.0b also confirms an almost prescient fear a BroadLogic exec expressed to me back in early 2005. He said, "People shouldn't be thinking that double or quadruple is good enough, because once you put these things in, they last a lot of time in the infrastructure—and you don't want to be ripping them out."
While doubling or quadrupling today's speeds might not be good enough in a few years, at least the cable industry, by adding the proper testing and interoperability safeguards, shouldn't have to invoke the forklift if they give 2.0b a go.