Comments on: Enterprise Computing: Is iSCSI The New Home Protocol?

By: Blessay: FCoE is JUST a transition Technology | My Etherealmind

Blessay: FCoE is JUST a transition Technology | My Etherealmind — Thu, 17 Dec 2009 21:29:21 +0000

[...] December 17, 2009 by Greg Ferro · Leave a Comment Hi! If you are new here, you might want to subscribe to the RSS feed to here more from me. Table of ContentsThe Four Common Falsehoods of StorageiSCSI is good enough, Customers buy good enoughNetworking isn’t reliable enoughStandards, Certification and Unicorn TearsFCoE is the ONLY choiceFCoE is a migration technologyNetworking isn’t exactly a saint hereiSCSI works when correctly designediSCSI needs to be updated tooFCoE is the only solutionFCoE suits Cisco and BrocadeWhat Networking still doesn’t haveWrapping It UpIt’s a common fallacy of Storage Experts is that FCoE will “rule the world” and that Storage equipment has some special magical properties that makes it different from any other technology. Chris Evans posted the most common fallacies iSCSI is the home protocol [...]

By: Howard Marks

Howard Marks — Tue, 01 Dec 2009 15:02:36 +0000

FCoE, unlike iSCSI, doesn’t and can’t run on Ethernet switches even Ethernet switches with DBC,CEE,DCE extensions unless those switches are specifically equipped with for FCoE. That means the switch vendor needs FC knowledge and only Cisco, Brocade and Qlogic are equipped there. This will keep FCoE as a much more expensive solution than iSCSI.

Why don’t FC shops run iSCSI:

1 – 2 solutions to the same problem is too many even if it has lower CapEx.

2 – We have storage guys have FC tools that don’t support iSCSI (in part a chicken and egg problem if big iSCSI nets existed the tool makers would add support)

3 – The network guys would be involved.

By: Alex Miroshnichenko

Alex Miroshnichenko — Mon, 30 Nov 2009 20:12:59 +0000

Would generally agree – I am actually a bit amazed that FibreChannel lasted that long..

However for home one of the biggest problems for iSCSI is the difficulty of sharing the storage devices at the file level – you would effectively need a cluster file system on all you clients which makes sharing across heterogeneous clients a problem. NAS devices have been doing for ages and at the performance in a typical home environment would be essentially the same as iSCSI..

By: Tweets that mention Enterprise Computing: Is iSCSI The New Home Protocol? « The Storage Architect -- Topsy.com

Fri, 27 Nov 2009 08:07:03 +0000

[...] This post was mentioned on Twitter by Chris M Evans, Emulex Links. Emulex Links said: Enterprise Computing: Is iSCSI The New Home Protocol? http://j.mp/8EyxXe [...]

By: Paul

Paul — Thu, 26 Nov 2009 17:19:03 +0000

Why do you think FCoE will triumph over iSCSI.

Surely when 10 GigE is prevalent, customers will prefer the much simpler and more flexible technology of iSCSI ?

EqualLogic is fast catching up on the feature set required by the vast majority of storage professionals, not to mention it’s considerably easier & quicker to setup and deploy !

By: Ewan

Ewan — Thu, 26 Nov 2009 13:10:34 +0000

I think most of the arguments you make above are true, but I think you’ve missed the biggest reason why iscsi never really made an impact in existing fibre channel environments – inertia

Enterprise It teams generally hate change, if they liked it they wouldn’t cope in the enterprise locked-down, don’t take a risk mentality.

In those kind of environments, anyone who suggested implementing iscsi alongside fibre channel would be viewed as slightly mad, so its much easier to justify the $1000+ per fibre channel network port

On the other side, I don’t see FCOE making an impact in smaller environments any time soon, 10Gb iscsi will be as you say easy, understood by the sole it administrator, and provide much higher performance than the disk array connecting to it, and usable on relatively cheap $1000 switches

By: Martin G

Martin G — Thu, 26 Nov 2009 10:44:58 +0000

Okay, but if 10GbE becomes consumer (which it will in time), why won’t FCoE make it into consumer devices? There are already software stacks out there!

BTW, although I think FCoE has a big future; mostly driven by storage traditionalists, I’m not convinced that iSCSI like rust will slowly eat away at FC. I used to feel that FCoE was really the future, I’m less convinced these days. In fact, I wish it would just go away but then that’s because I want to see more converged teams and that includes storage and network teams.