- David Passmore, Consultant and President of Netreference, a Boston-based research company, December 1998
Kicking off the 2002 fiscal year featured a major company reorganization. Cisco’s service provider line of business (SPLOB), among others, and its customer alignment was done away with and made way for a technology-based focus. Eleven new product groups were formed in the restructuring process. They included access, aggregation, Cisco IOS, Internet switching and services, Ethernet access, network management services, core routing, optical, storage, voice, and wireless. ATM was no longer considered a major product group, and in optical, Jayshree Ullal, replaced Carl Russo as the leader of the optical technology group.
Leading up to QWEST’s reduced spending was a critical meeting to identify where QWEST should spend its remaining capital dollars. QWEST had been on a spending spree, including modernization of its US WEST local exchange carrier, acquired, in part, for the purpose of filling the available bandwidth supply of its long distance operation. Most capital spending had been put on hold.
The topic of the planned meeting was about ATM, a technology that US WEST (now QWEST local) had insisted upon since 1998 . Even Cerent, to secure this Tier 1 service provider’s business, had to commit to developing ATM interfaces on the Cerent 454.
Todd adds, “We’re getting ready to go into this meeting with QWEST. It was a big Cisco account with five regions, a director, and the VP is taking a keen interest into what we’re doing [on the ‘454’].”
The internal preparation meeting was held with all of these executives. Looking back, Todd chuckled, “I’m surprised they even let me into the room.” As it turned out, Cisco’s ATM development leader was in the room too. Emotions were high.
Todd explained, “A few months before this meeting occurred, I [invited] some QWEST people down to [our Richardson facility] to see ATM integration on the 454 and talk about the ATM blades.” These on-board ATM switching blades would eliminate the need for standalone ATM switches.
Needless to say, Cisco’s ATM development team “went ballistic.” They saw this development on the successful ‘454’ platform as a threat to their standalone ATM box sales, an offering from their optical brethren that threatened the very existence of their ATM switches . Todd adds, “Not only did they not like me before, but now they also saw me as a threat and I was going to erode their bookings. They were so territorial. They wouldn’t talk to me and they instructed their guys not to talk to me.”
Todd was ostracized within Cisco, because of the fear of the ATM development team that their boxes could be replaced by integrated ATM switches on ‘454’ plug-in modules. “So we go into this meeting [in Denver]. We walk in and the guy who is Cisco’s ATM head and one of my peers, who realized he might lose some bookings, [are both standing there].”
The Optical Transport Business Unit, led by Terry Brown, had just cancelled the ATM program on the ‘454’, primarily, as it turns out, due to cost reasons, as Cisco demanded such high product margins. It became common knowledge that “it had been killed.”
They gloated. Todd was uneasy at their smugness. Referring to the cancelled ATM blades for the ‘454’, they said, “You guys just couldn’t pull it off.”
Todd recalls thinking that this attitude from coworkers is “not right.” All he could muster, “It’s pretty easy to kill something off.”
The customer meeting with QWEST went forward. Their senior vice president, Bill Adams, of architecture and engineering, the man Mike Peruse reported to was the host. Todd recalls his headcount check, “We’ve got a Cisco vice-president, a Cisco vice-president and general manager of the ATM business unit, every Cisco sales representative from the Denver office in the room.” All of the sales people were standing in the corners of the room. It was quite an imposing sight.
The QWEST executive walked in. He looked at all of the executives. Then he looked across the room at Todd. There was a pause as he took his seat.
He addressed the Cisco executives, “Why can’t the rest of your team treat us they way he does?”
Bill rested his gaze on Todd, referring to his relentless support. Todd was stunned, but recalls, “And there were just these looks of amazement on their faces.”
Bill said, “Optical people seem to get service provider. You guys need to learn from that acquisition.”
Talk about a big endorsement for the optical team within Cisco. “I was merely the face of what was optical and at the time, it felt good for me, but it was our team,” Todd said. “It was everything we did. It was the way we approached service provider. And it showed you at that time that Cisco still did not get it.” Indeed, even in 2002.
Todd reminisces, “We all used to talk about it. The optical guys [from Cerent] would look at Cisco’s IP guys and say, ‘You don’t get it. You don’t get Service Provider.’”
The whole idea of delivering products that supported the “five 9s” of reliability and quality, and then delivering on your commitments was crucial to customer’s placing repeat orders. The ATM meeting in Denver was just one of those events that really started to hammer home, for the larger Cisco organization, the importance of Service Provider’s needs.
The ATM development team and their sales folks didn’t recognize that Cerent’s acquisition served a two-fold purpose: “Cisco wanted into the service provider market,” Todd says, “and hey, if it’s through this new optical technology that we don’t have, that helps. That’s good.”
Shortly, thereafter, the ATM layer, as shown in this graphic, completely disappeared from the transport layer of the metropolitan and long-haul networks. One of Cisco’s major competitors, Nortel Networks, abandoned ATM too. This voice-based company recognized, in 2002, that the Frame Relay/ATM network was more expensive and complex to own and operate than optical Ethernet. Since service providers were beginning to move towards IP + Optical, Nortel noted that the complexity level using ATM-based solutions increased with the introduction of more advance packet services.
Meanwhile, Cisco’s ATM product portfolio, simply used by John Chambers’ executives to hedge their technology bets, began to wind down by 2004, with a series of end-of-life notices. Clearly, Cisco took their cue from QWEST too, as that company’s advanced planning folks projected in mid-2001, that by early 2003, ATM would no longer be part of their Metro Network Convergence plans.
Indeed, the premiere MSPP of its day, the ‘454,’ continued to chug along across the QWEST network, even without its once-planned ATM interfaces.
Todd simply smiled.
 Cisco’s John Chambers had agreed to Carl Russo, CEO of Cerent, and his demand that the Cerent optical sales force operate as a standalone entity for two years after the 1999 acquisition. Cisco honored their agreement, which lasted until late 2001, whereupon, both sales forces were collapsed into one. Todd Bremner sold $150 million of ‘454’ to QWEST in FY2001, but only $37 million in FY2002, a number, which included an expanded list of Cisco products.
 Don Brodigan, of US WEST’s Technology and Architecture Planning Department in Littleton, Colorado, stated, in July 1999, the ability to support both TDM and ATM on the next generation SONET platforms will be a requirement from US WEST moving forward. His colleague, Mike Logan, Data Applications Sales Manager for Strategic Internet Service Providers, identified an UUNET application, in October 1999, where that aggressive ISP wanted US WEST to hand them densely filled ATM OC-n connections into their mega-POPs, and have the telco groom all of the traffic closer to the customers. UUNET essentially wanted an all-ATM transport from the customer premise to their own mega-POPs. Note that no mention of Ethernet/IP was made at the time. In less than 2 years, QWEST’s tune would change, migrating to an “all-IP” evolution as “ATM withered on the vine.”
 Cisco’s BPX 8600 and MGX 8800 Series Multiservice Switches provided scalable sets of solutions to deliver ATM, Frame Relay, voice, circuit emulation, and IP services in small COs. The MGX 8800, for example, provided scalable switching from 1.2 Gbps to 180 Gbps switch capacity and from DS0 to OC-192/STM-64 (10 Gbps) interface capacity.