An Early DOCSIS Memoir: Roads Taken and Not

Most engineers have other things to do than write. Those who have the time and skill to draft their own conference papers or help marketing with white papers are rare finds. Fewer still are able – or willing – to shift from tech writing to historical narrative. But that’s what we see in Robert Cruickshank’s self-published e-book, The Souls of DOCSIS.

Part industry history and part personal memoir, this PDF document came out in 2015, roughly two decades after the events it narrates. Associated with a Cable Center initiative that celebrates the Data over Cable Service Interface Specification (DOCSIS) cable modem, this e-book is also an extended and gracious hat tip to a large number of friends, mentors, and colleagues.

Full disclosure: I know the author. In the book, he thanks me and several other trade journalists “for picking up the DOCSIS story 20 years ago, right about where this document leaves off.” That spirit of gratitude pervades “Souls.” Far removed from the stereotypical engineer, Cruickshank not only writes; he loves technology and people.

The Roads to DOCSIS

One benefit of reading history is the reminder of paths not taken. Cruickshank offers a few such object lessons early on. Remember Integrated Services Digital Network (ISDN)? Having encountered ISDN while at AT&T Labs, he thought it held “enormous promise.” Yet companies fought over it, misinterpreted the standards and never did interoperability testing, until too late. In the end, it became “road kill.”

Cruickshank also notes the wastefulness of permanent virtual circuits (PVCs) in ATM, which he had tested at the University of Colorado. (He pursued graduate work at UC after leaving AT&T and completing a stint in the US Air Force.) Another example comes from his early days at CableLabs. While managing responses to its first Telecommunications RFP, he found a surprising number of companies supporting next-generation digital loop carrier (NGDLC) technology, an approach he calls “absolutely brain-dead” and “a terrible use of shared network resources.”

But in that batch of responses, there were also a few “brilliant suggestions,” especially from LANcity and Com21. How the industry began moving toward a solution that, unlike the other examples, was both interoperable and efficient, takes up the rest of this story.

Asked by MSOs to provide a Consumer Reports-style comparison of existing cable modem technology, Cruickshank and his team leveraged available standards (IETF RFC 1242, in particular) and test equipment. Prominent among the latter was the DA-30C data network analyzer from Wandel and Goltermann. Other test tools included Network General’s Sniffer, Netcom Systems’ SmartBits and the OPNET Modeler and Planner.

DOCSIS components

One day, a call came in from Tony Werner, then CTO of TCI, who wanted to know how many modems an MSO could put on its upstream or downstream plant. That changed the game. It was time to move to more realistic RF testing scenarios. Enter the CableLabs engineering department.

In addition to CableLabs, the larger cable modem project now included MSO representatives, management consultants and vendor authors. Goaded by cable execs, the authors delivered a specification comprised of the following elements:

• A user-to-network interface (UNI) that leveraged multiple service identifiers (SIDs) and enabled application-by-application control;
• A physical (PHY) layer with tremendous potential for speeds;
• A media access and control (MAC) layer that leveraged Carrier Sense, Multiple Access with Collision Detection (CSMA/CD). (Cruickshank says the MAC was the spec’s “greatest accomplishment”)
• A Baseline Privacy Interface (BPI), which became the spec’s default security, in contrast to a removable security approach advocated by the parallel Multimedia Cable Network System Partners Ltd. (MCNS) initiative (another “road not taken” – and one Cruickshank labels as “insanity”);
• An Operational Support Systems Interface (OSSI) that Cruickshank says remains underused still.

Definitive History, Anyone?

Then the story ends, somewhat abruptly. Cruickshank labels this work Volume 1, so maybe another will come. Trade journalists did record much of the later history, as he acknowledges up front. Others, such as LANcity founder and CEO Rouzbeh Yassini, who played a key role in the next phase of DOCSIS, have also contributed to the written and oral record. Ongoing development of DOCSIS has relied upon vendor contributors, especially Cisco Cable Access CTO and Fellow John Chapman and other industry leaders, such as ARRIS CTO Tom Cloonan.

The story indeed continues, with DOCSIS 3.1 rolling out and distributed access architectures in trials. But if DOCSIS ever slows down enough for someone to write its definitive history, he or she will need to reckon with this short, exuberant and plainspoken account of its early days.