First, as we discuss in our Planning for FCoE article, the first wave of implementation does seem to be a rack by rack rollout with special focus on the virtualized server environment. However as this rollout continues we will have to address the issues with spanning tree protocol.
While this should not delay the initial rollouts, the latency issues that spanning tree has when a path fails and a new connection has to be made will become a greater issue. The subject is well beyond the scope of this blog and is the topic of an article we are working on. The good news is that the fibre channel community has an incredible track record on backward/forward compatibility so it is reasonable to expect that to continue.
Another important step for fibre in general and FCoE in specific will be to drive out the complexities associated with it. Part of this is merely legacy reputation, another is the complexities of dealing with any storage network, but part of it is earned. Fortunately the Fiber Channel Industry Association (FCIA) has an effort underway to address this. The plan is to offer the same performance and capabilities of traditional fibre but to keep the scale of the environment down. Think of it similar to how Intel defined the Netbook market to try to simplify the experience for the user.
This effort is critical for one important reason. As I have said in other "what protocol" blogs, users are most likely to select the protocol they are already familiar with. If fibre is not your first protocol for shared storage it is going to have a tougher time being the protocol of choice when your environment scales. A large step in obtaining this market will be convincing server manufacturers to start embedding FCoE on the mother board of their servers.
The good news is FCoE is moving forward and not a moment too soon. The industry does not stand still, we have seen an incredible resurgence in NFS thanks to virtualization and iSCSI continues to gain ground. New protocols like ATA over Ethernet (AoE) and even networked Serial Attached SCSI (SAS) could have a play in smaller systems. For the users all these protocol choices can be a bit overwhelming but in the end choice is good and competition will make everyone better.
Track us on Twitter: http://twitter.com/storageswiss
Subscribe to our RSS feed.
George Crump is lead analyst of Storage Switzerland, an IT analyst firm focused on the storage and virtualization segments. Find Storage Switzerland's disclosure statement here.