This caveat was very tough to achieve in a price-sensitive transport market where optical products were considered to be performing financially if they simply exceeded the 30 percent margin threshold, especially as they got closer to the access space. Cerent (Cisco’s early optical team) gambled that bandwidth would continue to pressure service providers to replace copper T1s with fiber, and fiber-rich buildings would continue to scale from 155 Mb/s to 600 or to 2,400 Mb/s rates. Technology available in the early 2000s provided much better product margins as bit rates increased, especially as component suppliers amped up sub-system integrations and achieved lower cost targets with increasing volumes.
But here’s the rub: The baby brother of the market-beater ‘454,’ the ‘327,’ was too much like the ‘454,’ owing to the engineers’ bid to reuse as much of the architecture and technology of the ‘454’ as possible. Furthermore, the ‘327’ ended up being too much like a telco product than a data product, even though its target was medium to large enterprises (businesses) in the access space.
Back row, from left to right:
Jeff Ogden, David Wong, Brian Acker, Chuck Dyer, John Diab, Thomas Steiner, Kamran Behbahani,
Angelo Purugganan, Don Turner
Front row, from left to right:
Kevin Hayward, Hao Tran, Kenny Choo, Soren Schuler, Lik Seng Lim, John Proctor
Missing from photo: Charles Chang, Luke Liong
As Scott Messenger, Cerent’s 454 and 327 product management leader recalls, “What happened was it drove the pricing up on the ‘327’ into an area that didn’t create enough of a price gap from the ‘454.’”
In that case, many customers simply bought the ‘454’ instead of the ‘327,’ seeing the ‘454’ as an insurance policy against increasing bandwidth needs in the future. Meanwhile, Cisco’s ‘327’ engineering team just couldn’t get the cost low enough to achieve the margins that Cisco wanted, critical for product pricing that allowed Cisco’s sales team to better compete against a myriad of new startup providers and reinvigorated larger equipment suppliers who were forced to innovate in the metropolitan optical transport space.
Scott adds, “We couldn’t drive the cost out of the product. It wasn’t chip technology, but modular subsystem electronics that we couldn’t move into the smaller ‘327’ form factor . . . We didn’t do a pizza box, but we did a smaller version of the ‘454.’ We didn’t get the densities out of it that we wanted to, in order to drop the price per port down to where we needed it [to better compete].”
But maybe the biggest thing that hampered the success of the ‘327’ was the DNA of the optical team itself, a group used to dealing with service providers demanding telco reliability and quoting Bellcore (Telcordia) specifications for redundancy and high performance in their network products. On top of that the environment was changing, as IP-based routing and switching products were not held to such high standards in the 2000-era, as the lines of telecom and datacom blurred.
Scott sees that today, “The second thing hampering us on the ‘327’ was redundancy. At that time we felt, and I supported what we did at that time, that taking a box that far out in the network, the expectation of five ‘9’s [availability] in a box becomes lower and lower. We were holding ourselves to a higher standard than we needed for ‘327’ reliability and we added in all of this telco level reliability and protection capability.”
Cisco’s optical team got caught in its own trap of building a telco-like product instead of a more data-like product as the solutions moved closer to the access space.
But most existing Cisco customers were not happy with the “high-priced” ‘327.’ Scott observes, “The 454:327 ratio was 10:1 and that 10:1 was weighted heavily on tw telecom. If it wasn’t for them, the ratio would be far worse . . . The ‘327’ was very price dependent . . . Some of the sales guys had to take the margins of the ‘327’ down to the 20’s, in order to create enough separation from the ‘454,’ depending on where they were selling the ‘454.’”
Dave Cesca, when he sold to the ‘327’s to tw telecom used this approach and kept the ‘454’ price right where he wanted it. The customer was happy too, as the right product was used for the right network application, with room for future growth.
As Scott is forced to conclude, “If nothing else, the ‘327’ was a margin protector for the ‘454.’”
But as every entrepreneur knows, the successful product has to win the hearts and minds of its customers.
Notes:
[1] Bell Communications Research was established by the Regional Bell Operating Companies upon their separation from AT&T. Bellcore, as it was nicknamed, was created in October 1983, as a centralized services organization resulting from the 1982 Modification of Final Judgment that broke up the Bell System into seven regional “Bells.” Later, this research and development organization was rebranded as Telcordia Technologies to distance itself from the “Bell” brand. This occurred subsequent to 1997, when Science Applications International Corporation (SAIC) acquired Telcordia. In 2000, the ‘454’ (Release 2.0) was the first Cisco product “OSMINE’d” by Telcordia. The cost: $3,250,000. Tens of millions of dollars would be spent on product certification (for suitability in the Bell network) by Cisco in the decade to follow. Then, in 2011, Ericsson acquired Telcordia and made it one of the Swedish company’s business units.
[2] tw telecom was acquired by Level 3 in 2015.
[3] According to Hui Lui, Cerent-Cisco’s hardware leader, “The XTC ASIC never happened because Cisco didn't want it. Otherwise, we could have had a cool ‘327.’ Without the XTC, ‘327’ couldn't be cost and power effective.”