Light Reading - Telecom News, Analysis, Events, and Research
Sign up for our Free Telecom Weekly Newsletter
Connect with us

Opinion  

100G Hits the Ground Running

February 19, 2010 | Simon Stanley | Comment (1)
   
 
no ratings

With a significant rollout starting in the data center and the first deployments of 100-Gbit/s transport in the long-haul network, 2010 is the year for 100G. Over the past eight years, many companies have made major investments in 40-Gbit/s technology with relatively little return, but with an eye on 100G. For some, that investment will now deliver a return, as 100G takes over for 40G as the highest-speed networking solution. The focus for 40G is shifting to delivering a cost-effective solution against multiple 10G and 100G options. Key elements in this transition are likely to be 40G serial interfaces.

The demand for 100G networking is growing quickly, driven by cloud computing, video, and other bandwidth-intensive applications. For some years, InfiniBand has delivered the fastest solution for high-performance computing connectivity, and this has continued through 2009. The introduction of 120-Gbit/s interfaces using active optical cables supporting 12 parallel links is a logical development from 40G InfiniBand. The specifications for 100G Ethernet (and 40G Ethernet) are now finalized, with ratification expected toward the middle of 2010. Several interface options included in the specifications build on work already completed for InfiniBand and 10G Ethernet.

Many carriers have been reluctant to upgrade 10G long-haul links to 40G, seeing this as an interim solution with limited payback. The first deployments of 100G equipment are delivering a new level of performance that can meet carriers' expectations for return on investment. At the moment, these 100G long-haul solutions are only available from leading telecom equipment manufacturers that have managed to invest in the required coherent receiver technology. This may signal a renaissance for Tier 1 vendors; however, the development of suitable modules and cost-effective silicon should restore the balance by 2013.

At this critical time, the latest issue of Heavy Reading Components Insider, "100 Gbit/s: New Technologies Deliver Bandwidth & Reach," analyzes the availability of 100-Gbit/s components, identifying the key advantages these technologies hold for data center managers, carriers, and equipment manufacturers. The report also reviews the availability of 40G solutions and profiles 25 leading vendors in these growing markets.

Active optical cables and transceivers for the 40G QSFP and 120G CXP format are available from four vendors. InfiniBand switch chipsets and network interface card silicon is available from Mellanox Technologies Ltd. (Nasdaq: MLNX). These same formats, QSFP and CXP, can be used for 40G Ethernet and 100G Ethernet in the data center. The imminent completion of the IEEE802.3ba 40G and 100G Ethernet specification is accelerating the introduction of these high-speed connections.

100G Ethernet over single-mode fiber is also coming quickly, with the introduction of CFP modules for both 2km and 10km applications. The CFP module format will also support Sonet/SDH and Optical Transport Network (OTN), making this ideal for client-side systems, such as routers. In these applications, pluggable CFP modules will quickly replace 300-pin fixed transponder modules.

Various vendors and carriers have used optical duobinary (ODB), differential phase-shift keying (DPSK), and differential quadrature phase-shift-keying (DQPSK) modulation schemes for 40G long-haul connections. Although these connections usually have similar equipment at both ends, there has been significant pressure within the industry to move toward a single solution. At 100G and transmission distances of 100 km or more, the technical challenges become much greater, and through work within the Optical Internetworking Forum (OIF) , the industry has coalesced around a single modulation scheme: DP-QPSK with coherent receiver. This solution requires advanced optical components, very fast analog-to-digital converters, and the latest digital signal processing techniques. As shown in the report, these elements are now becoming available for both 40G and 100G, and at least three manufacturers have already shipped systems.

A key rate for telecommunications with Ethernet, Sonet/SDH, OTN, and InfiniBand converging was 10G. The introduction of 40G interfaces has been relatively slow, with many interfaces costing more than four equivalent 10G interfaces. The latest developments at 100G show a clear trend toward this being the next widely deployed rate. 40G interfaces will continue to be used in place of quad 10G or where the extra bandwidth provided by 100G cannot be justified.

100G solutions require significant investment, but they also present a significant business opportunity for those vendors that have the right technology, and can deliver it at an attractive price.

— Simon Stanley, Analyst, Heavy Reading Components Insider


This report, 100 Gbit/s: New Technologies Deliver Bandwidth & Reach, is available as part of an annual subscription (6 bimonthly issues) to Heavy Reading Components Insider, priced at $1,295. Individual reports are available for $900. For more information, or to subscribe, please visit: www.lightreading.com/commchip.

Newest Comments First       Display in Chronological Order
HSalemi
User Ranking
Friday February 26, 2010 11:04:30 AM
no ratings
It will take longer than 2013 to get cost effective 100G long haul. Take a look at the time lines for 10G, first shipment around 1996! It has taken at least 10 years, a lot of R&D $ during the bubble, and now, we have good pricing for XFP and SFP+ modules! Todays investment R&D are very limited compared to 10G era, so, how can 100G development time lines be any shorter! Any vendor needs to get ROI back, therefore, for sometime, those modules are expensive and power hungry! Competitive pricing is beyound 2015!
The blogs and comments are the opinions only of the writers and do not reflect the views of Light Reading. They are no substitute for your own research and should not be relied upon for trading or any other purpose.