ffs, just move to unified fabric - 40Gbit / 100gbit Ethernet. We don't need Fibre Channel where we are going...
Storage networking biz Brocade has pulled a massive profit leap out of declining revenues thanks to a legal win against rival A10 Networks. Brocade banked a $119m profit on revenues of $536.6m in its latest quarter of 2013, ended 27 July. The year-ago profit was $43m, so that's a 177 per cent jump and a 153 per cent increase …
The lack of success for FCoE is down to two things in my experience:
1) control - who runs this network, the storage guys or the network guys?
2) sharing - the parts of the business dominated by storage consumers like databases don't want a bunch of ethernet traffic mucking up their latencies and the parts of the business that care only about the network faints from thinking about what terabytes of storage traffic will do to their network bandwidth. Doesn't matter how true this is, the perception of those who control/influence the purse strings does.
Since in many cases it isn't practical for the above reasons to share a common fabric, you still need two fabrics, which badly wounds the case for FCoE. While there would probably be some benefit to using 40Gb PHYs from ethernet instead of designing new 32Gb PHYs for FC to reduce the cost for both, I'm sure Brocade is avoiding that because they don't want to do anything to hasten the day when FCoE might succeed - because if/when it does, Cisco wins.
But with 100Gbps four links would be 400Gbps. Then you have 400Gbps Ethernet being possible with current technology. FC is just not going to survive. FC does have better latencies than some FCoE gear though, but higher speeds means lower serialization latencies. Queuing latencies could be worked on and bring Ethernet closer to FC.