Is the world's fastest router much ado about nothing?

In a high-profile announcement this week, networking giant Cisco took the wraps off its new CRS-3 router, which the company says is capable of handling up to 322-terabits of data per second. Among other platitude-laced marketing talk, the company says that the transfer rate is fast enough to transmit every movie ever made in "around four minutes."

While I harbor no doubts that the mind-boggling speed of the router is indeed attainable, or the engineering feat that made such a monstrous device possible, I couldn't help but get a sense that the announcement is nothing more than a marketing event which is getting milked for far more mileage than the actual, real-world impact of the new hardware.

For example, AT&T Labs President and CEO Keith Cambron was said to have chimed in during the webcast about how the company is confident that the CRS-3 will perform up to expectation. He cited how AT&T put the CRS-3 technology through its paces using a 100-gigabit traffic generator.

Now, I have no doubts at all pertaining to the veracity of Cambron's assertion, but the last time I checked, a terabit is equivalent to 1,000 gigabits of data. Put it another way, the 100-gigabit traffic generator could only pump out one-tenth of a terabit, or 0.031056% of the CRS-3 router's maximum capacity. Could someone kindly remind me why that made the news as a testament to the router's capabilities?

So what do competitors like Juniper Networks have to say about the CRS-3 technology? According to Alexander Wolfe, who is the editor-in-chief over at InformationWeek.com, Juniper has fired back at Cisco, accusing it of "funny math" in its speed claims and also of the implicit claim that Cisco got onto the 100 gigabit bandwagon first.

In an email to Wolfe, Mike Marcellin, who is the vice president of marketing of Juniper's Infrastructure Products Group and Junos Ready Software presented his company's take on Cisco's announcement. 

Marcellin noted that Juniper has been delivering 100 GB-capable systems since 2007, and also sought to downplay the 322-terabit capabilities of the CRS-3:

"The claim of 12 times the traffic capacity of the nearest competing system is based on a theoretical maximum of 72 interconnected CRS-3 chassis in order to achieve the 322Tbps total capacity. This will likely never be deployed in practice due to space, power and manageability realities."

As a former system administrator and IT manager, it is uncomfortably clear to me that configuring and maintaining 72 chassis would be an administrative nightmare.

Before I finish today's note, let me say that I have great respect for the engineering abilities of Cisco, and I am acutely cognizant of the tremendous advancement in network technologies that is the direct result of work done at the network giant. In a nutshell, the CRS-3 router is phenomenally fast hardware, and I can only guess at the difficulties inherent to spanning its network backplane across 72 chassis.

As tech editors and journalists however, I feel that we have an obligation to our readers to be more analytical when reporting on new technological products. While network experts and data center managers would undoubtedly have read between the lines within the first minute, the typical consumer--and even many IT professionals--would simply have taken any regurgitated marketing fluff at face value. Paul Mah (Twitter: @paulmah)