Comcast Exec Wants 1-Terabit Optical Standard

NEW YORK -- Packet-Optical Transport Evolution -- The optical networking industry needs to make a decisive choice now between 400-gigabit and 1-terabit options for the next generation of gear, says a top Comcast Corp. (Nasdaq: CMCSA, CMCSK) executive, who is opting for 1Tbit/s.

Shamim Akhtar, senior director of network architecture and technology at Comcast, made his first appearance at Light Reading's Packet-Optical Transport Expo Tuesday, and used his keynote platform to urge the industry not to replicate the lengthy 40G-versus-100G squabbles of the past, which can slow the pace at which new technology gets to market at prices that carriers and vendors can accept.

"Let's not repeat the 100G versus 40G debate -- let's stick with one, and my answer is 1-Terabit," said Akhtar, who throughout his speech stressed the similarities between Comcast's optical needs and challenges and those of its telco competitors.

Comcast is actively engaged with AT&T Inc. (NYSE: T), Verizon Communications Inc. (NYSE: VZ), Google (Nasdaq: GOOG), Facebook and others in a "100 G and Beyond" user group, Akhtar said. The group's initial goal is to get some unity on a 100Gbit/s client-side interface, and after that, they might tackle the 400G/1-Tbit question.

Comcast announced earlier this year that it has begun deploying 100G technology. (See AT&T, Comcast Go Live With 100G.)

"We didn't do Sonet, so OTN is not relevant to use," Akhtar said. "But we have a lot of similar problems to be solved. Instead of fix it twice, we can fix it once."

Comcast deployed a 40G national backbone in 2005 and according to Akhtar put the industry's first 100G production traffic on its backbone in March of 2008. Like AT&T and Verizon, Comcast is now looking for connectionless, directionless and contentionless ROADMs, as well "dense, less power-hungry equipment," Akhtar says. (See Verizon Readies 100G Launch in US.)

The Comcast exec did draw some key distinctions between his operations and those of large telcos, namely a smaller networking organization. But most of the requirements he laid out would sound very familiar to anyone in optical networking. They included better quality of experience for content delivery, an order of magnitude reduction in cost for delivering the coming exaflood of traffic, network restoration at the service level (and not the pipe level) and tighter integration at the logical level of packet and optical devices.

Akhtar also made a plea to eliminate the "alphabet soup" of standards development, saying that while standards groups evolved for a reason, the industry might be better served if there was one standards organization focused on carrier needs rather than the multiple groups -- the Institute of Electrical and Electronics Engineers Inc. (IEEE) , International Telecommunication Union, Standardization Sector (ITU-T) , Internet Engineering Task Force (IETF) and Optical Internetworking Forum (OIF) -- that operate today.

— Carol Wilson, Chief Editor, Events, Light Reading

DCITDave 12/5/2012 | 5:04:57 PM
re: Comcast Exec Wants 1-Terabit Optical Standard

I like the idea of declaring that 1Tbit/s is a priority and lining up all efforts to that eventual goal. That, it seems, could help focus the industry a bit more and avoid some squabbling in the process.

cnwedit 12/5/2012 | 5:04:57 PM
re: Comcast Exec Wants 1-Terabit Optical Standard

This strikes me as way too sensible to actually happen. There's got to be some major vendor with a stake in 400 G that can slow the whole process down and give us years of material.

kfritz 12/5/2012 | 5:04:47 PM
re: Comcast Exec Wants 1-Terabit Optical Standard

1 Tbps certainly seems the logical way to go due to greater needs to transport large amounts of data in a shorter amount of time and efficiently (be it medical centers or content providers).

The technology for Terabit looks feasible, but challenging.  For example, SerDes components are "still" in the 28 Gbps range.  40G electrical has been desired for awhile (2004 timeframe?), yet "we" still don't have it.  That said, multiple 28G or 40G lanes are going to be a large technological challenge to aggregate to 1 Tbps coherently - even if they do aggregate yet again to 100 G lanes.

The economics at this time doesn't look much better for 1 Tbps versus 400 Gbps..  This past year at the Ethernet Summit, the folks supplying the optical components do not seem to think 1 Tbps is economically viable yet.  My question is - how much of the investment for 100G has been recouped as of yet?  Can optical component manufacturers afford to put their profits on hold yet again to invest in another technology this soon (you could say.. "how could they not").

I have far more questions on this issue than I have answers.  I invite others to expand and refute my contributions to this thread.

Sign In