Skip to main content
Message Font: Serif | Sans-Serif
No. of Recommendations: 4
DD -

>10 Gig Fiberchannel and 10 Gig Ethernet are both >deliberately designed to steal as much technology >from each other and from SONET/SDH. This is good as >it brings products supporting both standards to >market sooner and means that they will be >interoperable to some extent.


>However 10 Gig Ethernet has one big advantage over >Fiberchannel that I can think of and that is that it >runs IP. FC is a data storage protocol and from what >I can tell a remarkably simple one at that. 10GE on >the otherhand can act as a storage protocol (using >NetApp style NASes) or it can be used to aggregate
>data traffic over the internet.

Factually incorrect. FC is a multiprotocol transport standard that supports storage (SCSI, IPI), network (IP), and clustering (VI) protocols natively. The IP protocol for FC is actually specified in an IETF RFC that has been ratified. We use IP/FC in our shop all the time - it is part of the core deliverables. (Also - to my knowledge, the single largest customer FC installation on the planet (based on node/port count) is in the far east, and THE primary protocol used by that customer is IP.) I have personally seen a multi-system pure FC cluster running all three protocols (FCP, IP, and VI), on switched FC fabric, and it handily whipped a conventional configuration (FC & Ethernet) by 2,200+ % in response time (interface clock rates don't even begin to tell the whole story at all here). Bottom line, to users who know the whole story here, the most powerful (and still commercially viable) IP networks can/will run on FC.

>I don't really consider the two protocols to be great >competitors actually because I think that FC will be >able to maintain its dominence in the back end >storage market with 10GE dominating the front end >data transport, but in terms of which one has the >potential to take the market from the other its more >likely that 10GE will take market from 10G-FC

Those who continue to view the diamond through the same 20 year old facet are missing over 90% of the stone. New models of applications and services will demand a new perspective, and those who stay in the box will miss it. The "back-end", "front-end" discussion will also pass. As for market share, with most of the installations that I am aware of, the ratio of FC ports to Enet ports is in the 2:1, 4:1, or even higher, range. For those just learning to walk before they run, the solutions tend to lean toward a 1:1 ratio. But that belies the true measure of cost effectiveness. Because a single storage subsystem may contain scores of FC chips (controllers, disks, etc) - hence development and manufacturing costs are spread more widely on a per unit basis. Also, unit costs is a deceiving metric. I much prefer a metric that factors in "costs per unit of work". (just for discussion) If GbE enjoyed a 25% cost advantage, yet FC yielded a 50% improvement in costs per unit of work, as a paying client, where are you likely to put your money to maximize your ROI??

This is only the tip of the iceberg... FC MANs and WANs are already showing up in production in the market, which will really help to drive volumes up, and unit costs down.

Just some thoughts.

Print the post  


What was Your Dumbest Investment?
Share it with us -- and learn from others' stories of flubs.
When Life Gives You Lemons
We all have had hardships and made poor decisions. The important thing is how we respond and grow. Read the story of a Fool who started from nothing, and looks to gain everything.
Contact Us
Contact Customer Service and other Fool departments here.
Work for Fools?
Winner of the Washingtonian great places to work, and Glassdoor #1 Company to Work For 2015! Have access to all of TMF's online and email products for FREE, and be paid for your contributions to TMF! Click the link and start your Fool career.