iSCSI dodges Fibre Channel over Ethernet noose Fibre Channel over Ethernet may not have taken off yet, but at least it's on the taxiway. That's thanks to NetApp's launch of the first native FCoE storage subsystem, converged network adapters (CNAs) from Emulex and QLogic, and Cisco's announcement of the first FCoE switch - the Connectrix NEX-5020, which is based on its Nexus …
"FCoE is kind of a grown-up iSCSI for large enterprises - for people who already use Fibre Channel and don't trust the kind of noddy TCP/IP-based technology used by small businesses."
That is an interesting analogy, which some might like to counter with a clip across the ear/shot across the bows to remind them as to who is Big Boss .... FCoE is the bastard problem child for large enterprises.
Where's the appeal?
Networking guys don't want to be in charge of networking storage.
Storage guys don't want the networking guys messing around with their storage networks
Directors of IT don't want to incur the switching costs from safe, stable (10+ years) Fibre-Channel.
This is not backwards compatibe with FC4/FC2/FC1. FC8 is backwards compatible.
Businesses leveraging iSCSI - not just SMBs, but real enterprises running real applications like Exchange, VMware, SQL, etc - do not want to move to a proprietary dedicated and more expensive topology when they have one that already works very well.
Who's pushing this and what's the value proposition? Answers in 3 sentences or less otherwise you disqualify yourself :)
> who's pushing this?
FC vendors.
> value prop.
Fibre-channel has already over-shot the requirements of 99% of customers with FC4. This enables you to buy more stuff and here is a new, complex standard that's better because it's newer and doesn't have SCSI in the name since that sounds old.
Who else remembers how ATM was going to take over the desktop? There was going to be ATM in every switch, every closet, every enterprise. Was it fast and reliable? Yes! So what happened? A plucky little contender called Ethernet came along (thick-net/ vampire taps/BNC-style) and although it wasn't nearly as reliable or as fast as ATM it was way cheaper and more importantly, it was "good enough".
10-GigE is here and 40-GigE is just around the corner (and 100-GigE not far behind). iSCSI over these speeds may not be as feature complete as FCoE on the same pipe, but it *will* be cheaper and it *will* be "good enough" for all but the most demanding enterprises (or those with the "well, we've already spent so much on this we might as well keep doing it that way").
SAN guys take heed, learn ethernet or lose. FCoE is a last gap for a dwindling market. Why Cisco is hitching up to that train I have no idea.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
