By Contributing Editor Stephen Hardy, Lightwave
3:20 PM PST: In a break at the OSA/Lightwave Executive Forum. I'll have an article later on the 40G/100G components and subsystems panel I moderated this morning. Meanwhile, here are a few select snippets from other panels and speeches:
* Google's Vinton Cerf admitted that he didn't envision Twitter when he "co-invented" the Internet. He says he asked the head of Twitter whether his title was "chief twit"...Cerf said that while fiber makes the most sense for the majority of Google's FTTH trials, the search engine giant is pondering the use of microwave at some edges of its networks, principally where laying fiber would be too expensive...
* Ovum's Dana Cooperson said that the WAN DWDM 40/100G CAGR between 2008-2014 will be approximately 79%. Meanwhile, the CAGR for 40GbE/100GbE transceivers will be 200% 2009-2015. There will be more than 2.5 million 40/100GbE ports shipped in 2015. Meanwhile, she thinks it will take 8-10 years to get to 100G to 6x cost for 10x the bandwidth compared to 10G...
* Bikash Koley, network architecture at Google, reports that there's no incentive to install 40GbE or 100GbE until the 20x power penalty versus 10GbE typical for bleeding edge technology approaches parity. He also said that, since Google already is running 800 Gbps per fiber, the choice of 100GbE versus 40GbE is meaningless when the standards only assume one interface per fiber. He'd like to see a rate adaptive version of 100GbE, similar to what's available in certain flavors of DSL (albeit at lower bit rates)...
* Hans-Juergen Schmidtke of Nokia Siemens Networks, says that 40G/100G optics costs will come down once “overlooked manufacturing innovations (automation) become effective." So stop overlooking automation, you components and subsystems vendors! Meanwhile, he also offered that there's plenty of opportunity to innovate and differentiate at the component/subsystem level.
Gotta go -- the next session is starting...
SOURCE:The Lightwave Blog