Comments on: The Anti-FCoE Sentiment

By: Niall McGee

Niall McGee — Thu, 22 May 2008 19:35:31 +0000

Dante,With regard to your point about alternative topology selections for Ethernet, that looks real interesting. Will be much better more interesting when (a) Cisco have decided which horse they’re backing (IETF or TRILL) and/or (b) one of the two is standardised!Your point that … the large majority of storage traffic is local, …”" is valid, but “”local”" != “”L2 adjacent”". It’s perfectly common to have storage >=1 L3 hop away, and such a topology gives substantial increases in flexibility.Having said this, I do see a position for FCoE. I’m much more excited about the other prospects opened up for IPoE traffic by DCE, though!Rgds,Niall.”

0 likes

By: Greg Ferro

Greg Ferro — Thu, 22 May 2008 19:35:31 +0000

I am not anti-FCoE, I am anti-Fibrechannel. After Cisco led the market by being the one of the first companies to release an iSCSI to Fibrechannel router (the SN5420 from memory) in 2000 or so, the takeup of iSCSI has been slow. I do feel that Fibrechannel has delivered far too little for the price we have paid. I have posted a response at http://etherealmind.com.“

0 likes

By: Anton Kolomyeytsev

Anton Kolomyeytsev — Thu, 22 May 2008 19:35:31 +0000

FCoE was invented to keep whole FC going. There’s absolutely no other sense in it. Except what Greg pointed to (and he’s 200% correct) there’s another small little thing”" virtually nobody talks about. Nearly all of the modern NIC ASICs have TCP optimized data transfer engines. So if working on channel level (where Ethernet frames live) will give much worser results (both latency & thruoutput) compared to the TCP. My company is selling Windows AoE (ATA-over-Ethernet) initiator and we’ve already hit what’s FCoE guys will see when FCoE will be in the production phase…Regards,Anton KolomyeytsevCEO, Rocket Division Software”

0 likes