Monday, August 20, 2007

FCoE ready for ASICs

Kudos to Brocade and Cisco for agreeing on a single frame format for Fibre Channel over Ethernet. They struck a compromise that moves the standard forward, although it probably requires both companies to rework any ASICs they have in development.

Now chip development can start in earnest for adapters that will handle both Ethernet networking and Fibre Channel storage. But there is plenty of software work yet to be done to finish the spec, and no final standard is expected until April 2008.

Ah, to live in a simpler world where everything rides over Ethernet. We are stepping closer to such a world—even as the Fibre Channel folks gear up to leapfrog good old Ethernet yet again. More on that later.

1 comment:

Anonymous said...

While a single frame format is a good step, it is only a small fraction of the functional semantics that need to be specified. Your article touches upon only a sliver of what remains to make FCoE a reality. Keep in mind that it took 10 years for FC to get its act together and IB-FC gateway products which have been specified for a number of years, still have performance and quality challenges in the market though those are being addressed.

FCoE isn't a panacea. While it can lead to a unified fabric type being emitted from servers, it is quite likely customers will deploy separate instances per usage type - one Ethernet instance for storage and one for network applications. This will be just the first step too. The whole management quagmire will take a great deal of time to sort out and then translate into a quality offering.

One question you have not touched upon is will the major Ethernet NIC providers - Intel and Broadcom - push FCoE into their NICs or will they rely on software based initiators (much of what has been done with iSCSI due to the nascent nature of the market). If these two companies don't translate FCoE into hardware early in any potential adoption cycle - that is a significant challenge - then customers may shy away and take the same adoption process as done in iSCSI.

The other point you don't cover is that FCoE relies upon IEEE DCE being developed and deployed. That effort has seen some set backs. Not insurmountable but given how long it will take to get some aspects defined such as congestion management, the prognosis for FCoE's rapid adoption isn't that good. Best bet for most people is the earliest some customers might do more than play with the technology is late 2009 or perhaps well into 2010. Don't fall for the marketing hype from those providing gateways - perhaps that is their only goal which is to create lots of islands of functionality that require a big company like Cisco or Brocade to sort out and make these islands operate as customers desire and are likely to be disappointed when they see the real cost and complexity of FCoE.

It is too bad that instead of pushing the same 30+ year old SCSI paradigm which has been permuted again and again with significant fabric complexity, high costs, and cusomter frustration, people did not take the time to really step back and question is there a better way to do storage that just does not re-invent the wheel and add cost and complexity. FC will eventually die like any technology will but please don't view its rebirth on top of Ethernet as a major step forward. It isn't about improving storage solutions but about enabling companies like Cisco and Brocade to sell more complex management packages and create more customer control points. What a great world we live in these days.

 
interconnects