No, in the year 2000, this was not true. Edwards must’ve been comparing the best of long-haul DWDM against the decades old, high-priced traditional SONET platforms. It was well known by the fall of 2000 that the Cerent 454 (or Cisco ONS15454) in its OC-48 configuration cost less than legacy OC-12 systems from vendors such as Nortel, Lucent, Alcatel, and Fujitsu.
Not true, either. A Cerent 454 OC-12 or OC-48 configuration cost less than this. Furthermore, a forklift upgrade was not needed to increase capacity of a Cerent 454 from OC-12 to OC-48, a 400-percent increase in capacity, or even upgrade to OC-192, a 1,600-percent increase in capacity. Simple plug-in trade outs, performed in-service, accomplished this upgrade. The time to install a Cerent 454 was hours, not days for traditional SONET gear, and the time to provision a circuit only seconds not weeks.
Edwards persisted in charactering SONET as “designed and optimized for voice, expensive and inefficient for data-centric applications, does not scale well and is costly to upgrade.” Cerent turned this conventional wisdom on its head in early 1998, when customers recognized the startups’ value proposition and placed orders for the Cerent 454.
Cerent’s product, of course, embraced voice-based traffic (TDM), but it also efficiently enveloped data-centric traffic. By simply equipping one of its many multi-service interfaces, the Cerent 454 could transport IP packets or ATM cells [2]. On top of that, the product sported an integrated 3/1 and 3/3 cross-connect capability (DACS), which obviated the need for standalone electrical cross-connects at most of a telco’s central offices. The ability to scale in terms of service interfaces, bandwidth management, and optical bit rate made next generation optical transport not only desirable, but an MSPP evolution that persisted for more than a decade. Cerent’s MSPP definition delayed the introduction of Metro DWDM on a large scale for almost as long.
While Ciena eroded Nortel and Lucent’s market share in the long-haul market, Cerent did so in metropolitan networks, in spite of Edwards’ prognostications.
Thank goodness for upstart startups.
[1] According to Sterling Perrin, a former analyst at IDC, Morris Edwards was not an analyst with the firm. That may explain why his October 2000 article – DWDM heralds abundance of bandwidth – was out of touch with reality. Rather, Edwards worked for the consulting side of IDC as a project manager. These consultants, I was told, generally “pulled together research from various IDC analysts for white paper projects, so they were not experts in particular areas.” This may be why consultants like Edwards generally weren’t quoted in the press. Sterling related that it was quite a surprise to him that Communications News would’ve quoted him, let alone published an article on optical transport by him.
[2] Cerent, within Cisco, developed three ATM blades for the ‘454’ platform, a need articulated by the Regional Bell Operating Companies (RBOCs). Ultimately, Terry Brown, Cisco’s GM for optical, pulled the plug on the ATM development before these three multi-service interfaces reached the field. By then, even the ATM stalwarts, the RBOCs, conceded that packet switching and Ethernet transport was well on its way to replacing circuit switching and ATM, in particular.