An additional interface that could capture an incremental $125 million in business was a no-brainer. MSOs, such as Cablevision Systems, were engaged and a product specification was developed – to directly transport analog video across Cerent’s optical fiber network encased in a 155 Mb/s (STS-3c) envelope. Bandwidth was becoming cheap, so sending uncompressed video across optical transport networks further lowered overall transmissions costs by cutting out expensive compression equipment. After Cablevision visited Petaluma and convinced Carl Russo that there was a significant market for analog video over SONET, the project was born.
Bob Bortolotto, one of Cerent’s key hardware leaders, initially led the engineering team, which included Pat Garvin [1], to develop video cards that would overcome the noise problem and others. The initial design featured a single card in support of four video channels (RF and audio). Each channel would be independently selectable for baseband video and audio or an Intermediate Frequency (IF).
He continues, “I had suggested that we create a separate box to do the video conversion from analog to digital and then pass that digitally to the ‘454.’ I was overruled on that suggestion because part of the selling feature of the ‘454’ was that it was a single unit [supporting] multiple services.”
Pat was correct. If analog video could not be integrated into the Cerent 454 directly, there was no point in producing video plug-in options for the cable operators. Standalone boxes produced by Nortel or ABL had been on the market for years and were already being used in a number of Cerent bids to support digital video transport. Bringing another add-on video conversion box, late to market, was not an option for Cerent.
At the kickoff meeting for the product release that would support the video card for the ‘454,’ Pat presented the high-risk nature of the video project. As he says, “No one seemed to listen.”
Tom Randstrom, Cerent’s Marketing Manager recalls other difficulties with the video project, “As development progressed, the ability to support either type of video interface on any input on a single card was lost, which would have added ‘some’ cost in mixed channel networks. The program also continued to slip its schedule, neither of which was a deal killer, in my analysis.”
Pat counters this assertion, At least “we could do baseband video and audio with confidence and I offered that as a solution. No one took it. I do know that Cisco tried selling a similar product years later.”
But there is more to it than that. Architectures of a CATV network demand the distribution of video services from the headend to the home be encrypted so that the operators’ video programming cannot be viewed by unauthorized users (i.e. non-paying customers). The video signal gets encrypted at the headend by an encoder and decrypted at the home using a set top box.
Cerent’s sales team was not interested in excuses and the pressure to deliver was intense, especially after the decision was made to pull the plug on the video project.
Of course, Cerent, like all startup companies had a finite number of engineering resources and something had to give. Furthermore, Cerent’s sales team could not hire enough qualified people to catch a lot of the business being thrown its way for the base SONET applications.
Tom notes, “I think the real killer of the video program was the potential growth from other applications and customers. When we started the video project, Cerent was sub-$10M in sales. When the video project was killed, the growth in the company was starting to be realized; which was much bigger than the $125M total video market we were chasing.”
Tom is right. Within a year of the cancellation of Cerent’s video transport foray, the company hit a billion-dollar run rate. Even so, says Tom, “I still believe, if we would have released the video card, even with its restricted provisioning capability, it would have stomped ADC and Scientific-Atlanta's video solutions and brought the first integrated video card into any MSPP platform.”
As Tom says, “We can all dream!”
Pat is more pragmatic about the situation in 1999, “Around that time, we were bought by Cisco. Through all of the excitement of the acquisition, no one within Cisco seemed to care about adding analog video to the platform. I remember traveling to Cisco’s San Jose campus with Scott Messenger to present our video solution to their systems engineering staff. They weren’t interested in analog video over SONET. They scoffed at us and said we should have developed MPEG2 over SONET instead. But an MPEG2 solution already existed since most MPEG2 video encoders transmitted digitally compressed video over ATM-AAL5 over OC-3c. In our case, we were building analog video for MSOs who had committed to buying it.”
The irony of this is a few years later Cisco would get into the video business in an even bigger way by acquiring Scientific-Atlanta and secure this acquisition for the same dollar amount as the Internet giant’s $6.9 billion outlay for Cerent.
Cerent always acted boldly, even if a project did not make it to market. Cerent’s management team pushed the boundaries by attempting to add analog video transport as a multi-service interface to the ‘454’, but noise doomed this potential offering. The challenge of noise, in the form of jitter, almost scuttled the 10Gb/s optics option for the ‘454’ too, but in this case, more time and money was provided by Cisco and that solution made it to market.
But that is a story for another day.
[1] At the outset, Pat was the only engineer with any sort of video experience. He had worked for General Instrument for a decade doing digital ASIC design and some analog video design prior to joining Cerent on December 9, 1997. The team on the video development soon included Mark Baldassari, John Proctor, Mark Bennett, and Jason Reese who “made huge contributions. Without them,” Pat says, “we would never have gotten as far as we did. We even brought specialists from FPGA companies Xilinx and Altera to make the product a reality. Together, we were able to put together a system that passed video through the ‘454.’” But that was only a lab demonstration for Cablevision Systems.
[2] Tom worked at both ANTEC and Nortel on video projects prior to joining Cerent in 1999, to support Doug Juers and the MSO sales team, “We had to overcome this same obstacle at ANTEC, back before my time at Nortel, when trying to use an early 2-channel video to DS-3 CODEC (with pre-standards compression) to transport video over SONET. Our resident engineering guru, Jim Farmer, designed a device that would allow the encryption to be distributed without the need for an encoder at each secondary headend. Unfortunately, our partner company developing the CODEC could never deliver a stable product and the program was subsequently cancelled.”