100Gbps per lambda workshop is most timely

By Dale Murray


Questions on 100GbE on a single lambda? We’ve got plenty. Straightforward NRZ signaling probably won’t get us there.  So what type of advanced modulation would be the best choice?  How do we write a standard for advanced modulation? Mega datacenters are challenged for power consumption and we are approaching a limit on power consumption in a rack.  So how can we do 100GbE on a single lambda with power efficiency in a high-density line card?  Most importantly, how can it be done at the economic cost that Ethernet always delivers in the datacenter?  Adopting 100Gbps solutions from the telecom industry is unlikely to get us a solution the datacenter operator will be able to afford. 

Finding the investment required to develop, test and manufacture 100GbE per lambda modules will be another challenge.  How much investment will be needed and how do the investments get focused on the most attractive solution? Can we avoid fragmenting the investments across competing technologies?  When is the investment needed and where will it come from?

Relax, you may be thinking. We all know we will need 100GbE on a single lambda someday, but we’ve got plenty of time to work it all out, right?  That was my first gut reaction.  But just last week the IEEE 802.3bs Task Force on 400GbE met for the first time and straw polls indicated that 100Gb per lambda may become a favored solution.  And last proposals for 400GbE are due this November! 

So the time is now to get more industry discussion underway.  The OIDA/Ethernet Alliance ‘100GbE per Lambda for Data Center Workshop’ on June 12-13 could not be timelier.  Make sure you get your voice heard and hear what others are thinking.  Don’t put off registering today!