“What to do?” was the question on the minds of many in the telecom industry at the turn of the millennium.
Go for a 40 Gbps optical transport system (based on SONET/SDH) in the metro?
OR
Continue to invest in Dense Wavelength Division Multiplexing (DWDM) technology for metropolitan networks?
This decision faced Jayshree Ullal and Cisco’s optical team, partly, in light of a February 2002 report released by the US Department of Commerce, A Nation Online: How Americans are Expanding Their Use of the Internet.
The Federal Government’s analysis underscored the need for more capacity. Some two million new Internet users per month were being added to service providers’ networks in spite of the dot.com bust and telecom meltdown.
“In September 2001, 143 million Americans were using the Internet – an increase of 26 million in 12 months,” the report stated. “Internet use is increasing for people regardless of income, education, age, race, ethnicity or gender.”
Traditional TDM-based networks (wired phone lines) were the predominant medium of choice over newly arriving DSL capabilities (offered by telcos) and cable modems (offered by cablecos). At the time, eighty percent of Americans accessed the Internet using dial-up service made possible by Internet Service Providers such as America OnLine (AOL), with users typically using a newly installed second or third phone line.
According to the Commerce Department, “Between August 2000 and September 2001, residential use of high-speed, broadband service doubled.”
Although the demand for bandwidth existed and appeared to be expanding, excess capacity couldn’t be consumed fast enough, which is why many new metropolitan service providers merged or failed during 2002 through 2004.
What Cerent believed all along, and Terry Brown and I (Rob Koslowsky) were advocates of this belief as early as September 1998, was that by augmenting the Cerent 454 OC-48 functionality with OC-192 capability for metro applications, this more cost-effective alternative would delay the introduction of competing Metro WDM-based products.
Go for a 40 Gbps optical transport system (based on SONET/SDH) in the metro?
OR
Continue to invest in Dense Wavelength Division Multiplexing (DWDM) technology for metropolitan networks?
This decision faced Jayshree Ullal and Cisco’s optical team, partly, in light of a February 2002 report released by the US Department of Commerce, A Nation Online: How Americans are Expanding Their Use of the Internet.
The Federal Government’s analysis underscored the need for more capacity. Some two million new Internet users per month were being added to service providers’ networks in spite of the dot.com bust and telecom meltdown.
“In September 2001, 143 million Americans were using the Internet – an increase of 26 million in 12 months,” the report stated. “Internet use is increasing for people regardless of income, education, age, race, ethnicity or gender.”
Traditional TDM-based networks (wired phone lines) were the predominant medium of choice over newly arriving DSL capabilities (offered by telcos) and cable modems (offered by cablecos). At the time, eighty percent of Americans accessed the Internet using dial-up service made possible by Internet Service Providers such as America OnLine (AOL), with users typically using a newly installed second or third phone line.
According to the Commerce Department, “Between August 2000 and September 2001, residential use of high-speed, broadband service doubled.”
Although the demand for bandwidth existed and appeared to be expanding, excess capacity couldn’t be consumed fast enough, which is why many new metropolitan service providers merged or failed during 2002 through 2004.
What Cerent believed all along, and Terry Brown and I (Rob Koslowsky) were advocates of this belief as early as September 1998, was that by augmenting the Cerent 454 OC-48 functionality with OC-192 capability for metro applications, this more cost-effective alternative would delay the introduction of competing Metro WDM-based products.
We were spot on in our analysis, as the incorporation of WDM technology into the metropolitan segments occurred some three years later. Service providers were more comfortable with deploying SONET/SDH into metro and even regional networks in the early 2000s than adopting cutting edge WDM offerings, which they felt, would force them to change how they operate and maintain their installed base of optical transport networks.
This delayed technology introduction (i.e., slowing widespread migration from SONET to WDM in the metro) ultimately had the effect of killing the need for OC-768 capability (40 Gbps SONET solutions for the U.S.).
Hui Lui’s team, at Cisco, had already developed an OC-768-capable switch matrix (with a SXC ASIC), but not the associated BTC ASIC. Hui Liu recalls, “We had the technology and know-how to do the OC-768. But that didn't happen because of DWDM.”
This delayed technology introduction (i.e., slowing widespread migration from SONET to WDM in the metro) ultimately had the effect of killing the need for OC-768 capability (40 Gbps SONET solutions for the U.S.).
Hui Lui’s team, at Cisco, had already developed an OC-768-capable switch matrix (with a SXC ASIC), but not the associated BTC ASIC. Hui Liu recalls, “We had the technology and know-how to do the OC-768. But that didn't happen because of DWDM.”
Jayshree was at the helm when this 40 Gbps TDM-based technology was canned and the ‘454’ platform was split into two development streams:
She told me, in 2013, “The Cerent team and the Pirelli team of technologists had to develop a strategy [inclusive of the ‘454’ MSPP platform evolution]. Saying IP + Optical is a nice high-level strategy, but then how does it translate into products. If we just played into the IP over DWDM strategy of the Ciena’s of the world, all we would have done was mimicry.”
Jayshree continued, “But if you carefully position and lead off with your strengths:
- the provisioning aspects of the Cerent platform, and
- the IP and data aspects of Cisco,
you can successfully create a new category and lead.”
And here is where she brought in the collaborative aspect of the Petaluma-based and Monza, Italy-based optical expertise, “If you say that’s still important but I [also] need to do specialized DWDM and make sure it maps into that, which is what Multi-Service Transport Platform (MSTP) became. I think it was some brilliant positioning to get that segmentation. Then we defined the market and our customer base and how to lead in it, rather than just copying an existing, classical definition.”
OC-768 development was canned. Indeed, SONET (and its international SDH complement) would eventually fade and be replaced by WDM-based solutions featuring ROADMs and then even newer optical standards leading to Packet-Optical Transport solutions. Cisco played a big role here too, even as Jayshree moved on.
After Jayshree’s departure, David Friedman recalls how Cisco senior management altered the trajectory of the optical team, in 2006, “They split the DWDM [MSTP] team away from the SONET/SDH [MSPP] group. DWDM was viewed as growth and it was needed in the CRS product and that’s when [Cisco] did the ‘IP over DWDM’ blades . . .” [1]
According to another early 2002 study by McKinsey & Company, “the overall market for optical networking equipment went from $28.5 billion in 2000 to $19.7 billion in 2001,” reflecting the reality of the telecom meltdown. However, according to the same study, “the market for next-generation optical equipment actually increased from $3.5 billion to $4.9 billion in the same time frame” indicating the desire for newer solutions such as the Cerent-Cisco ‘454.’ This thirty percent market growth in a very tough economy reflected the appeal of next-generation optical transport equipment that delivered a superior return on investment.
What happened?
OC-768 never materialized, and instead, WDM-based solutions prevailed, just as Terry and I thought they eventually would. Jayshree executed on that DWDM vision to maximize sales of the ‘454’ for the evolving optical market [2].
[1] “What happened, when I was in Atlanta,” David Friedman recalls, “Scientific Atlanta was just acquired and they had some SONET products and so the Cisco executives decided to keep the optical transport stuff with the cable stuff and put the MSTP team into the CRBU – Core Routing Business Unit – under Mr. Kelly Ahuja . . . It was no fun, and we were the red-headed step child in CRBU.” This was the low point for optical within Cisco; things improved when Surya Panditi subsequently took over the optical helm.
[2] By 2010, the traditional MSPP had been superseded by the MSTP incarnation of the ‘454.’ The baby brother, the ‘327,’ and its subsequent larger sibling, the ‘600,’ both fizzled in the market, in favor of the transformed ‘454’ becoming an MSTP. The graphic below, shows that a new set of Cisco-spawned siblings was introduced with the product’s tenth release to take advantage of both the low end (M2 and M6) and high end (M15) of the growing MSTP-ROADM optical transport market.
- The SONET/SDH-based MSPP offering (primarily in a sustaining role), and
- The DWDM-based MSTP offering (with strong collaboration using the optics expertise from Cisco’s Pirelli acquisition)
She told me, in 2013, “The Cerent team and the Pirelli team of technologists had to develop a strategy [inclusive of the ‘454’ MSPP platform evolution]. Saying IP + Optical is a nice high-level strategy, but then how does it translate into products. If we just played into the IP over DWDM strategy of the Ciena’s of the world, all we would have done was mimicry.”
Jayshree continued, “But if you carefully position and lead off with your strengths:
- the provisioning aspects of the Cerent platform, and
- the IP and data aspects of Cisco,
you can successfully create a new category and lead.”
And here is where she brought in the collaborative aspect of the Petaluma-based and Monza, Italy-based optical expertise, “If you say that’s still important but I [also] need to do specialized DWDM and make sure it maps into that, which is what Multi-Service Transport Platform (MSTP) became. I think it was some brilliant positioning to get that segmentation. Then we defined the market and our customer base and how to lead in it, rather than just copying an existing, classical definition.”
OC-768 development was canned. Indeed, SONET (and its international SDH complement) would eventually fade and be replaced by WDM-based solutions featuring ROADMs and then even newer optical standards leading to Packet-Optical Transport solutions. Cisco played a big role here too, even as Jayshree moved on.
After Jayshree’s departure, David Friedman recalls how Cisco senior management altered the trajectory of the optical team, in 2006, “They split the DWDM [MSTP] team away from the SONET/SDH [MSPP] group. DWDM was viewed as growth and it was needed in the CRS product and that’s when [Cisco] did the ‘IP over DWDM’ blades . . .” [1]
According to another early 2002 study by McKinsey & Company, “the overall market for optical networking equipment went from $28.5 billion in 2000 to $19.7 billion in 2001,” reflecting the reality of the telecom meltdown. However, according to the same study, “the market for next-generation optical equipment actually increased from $3.5 billion to $4.9 billion in the same time frame” indicating the desire for newer solutions such as the Cerent-Cisco ‘454.’ This thirty percent market growth in a very tough economy reflected the appeal of next-generation optical transport equipment that delivered a superior return on investment.
What happened?
OC-768 never materialized, and instead, WDM-based solutions prevailed, just as Terry and I thought they eventually would. Jayshree executed on that DWDM vision to maximize sales of the ‘454’ for the evolving optical market [2].
[1] “What happened, when I was in Atlanta,” David Friedman recalls, “Scientific Atlanta was just acquired and they had some SONET products and so the Cisco executives decided to keep the optical transport stuff with the cable stuff and put the MSTP team into the CRBU – Core Routing Business Unit – under Mr. Kelly Ahuja . . . It was no fun, and we were the red-headed step child in CRBU.” This was the low point for optical within Cisco; things improved when Surya Panditi subsequently took over the optical helm.
[2] By 2010, the traditional MSPP had been superseded by the MSTP incarnation of the ‘454.’ The baby brother, the ‘327,’ and its subsequent larger sibling, the ‘600,’ both fizzled in the market, in favor of the transformed ‘454’ becoming an MSTP. The graphic below, shows that a new set of Cisco-spawned siblings was introduced with the product’s tenth release to take advantage of both the low end (M2 and M6) and high end (M15) of the growing MSTP-ROADM optical transport market.