The likelihood is that storage arrays will be linked to the Nexus in this scheme by Ethernet and not Fibre Channel. MDS9000 Fibre Channel fabric switches won’t appear in this setup because including Fibre Channel will be the very antithesis of a unified fabric.
Let me give you my perspective on this. Fibre Channel is all but the antithesis of a unified fabric, it’s part and parcel of it and it’s here to stay, as independent technology or co-existing with other connectivity options, including FCoE. Which by definition is Fibre Channel, simply carried over Ethernet. And MDS 9000 Family, Cisco Storage Area Networks, is a critical component of Cisco Unified Fabric strategy to support Fibre Channel, Storage Services and, in the future, FCoE connectivity to complete the unification of SAN and LAN. I talk a lot to customers, journalists, analysts and the question about the disruptive introduction of FCoE does come up all the time. And it comes up in a consistent variety (what an oxymoron!) of ways. Is FCoE going to kill Fibre Channel (and hence is Cisco not investing anymore in MDS 9000 product line)? Is FCoE going to kill iSCSI? Or is iSCSI going to kill FCoE (and maybe Fibre Channel too, why not?) I typically split my answer in two parts: the first about technology, the second about Cisco approach to technologies in general. It helps a lot to clarify how the future (and Cisco commitment) of the storage industry will shape.
From a technology standpoint I personally think that nothing will kill (what a strong word btw) anything. I see iSCSI, FCoE and Fibre Channel playing, today and in the future, different roles, addressing different market segments and solving different customers’ problems. Let me try to simplify as much as possible the situation: -- iSCSI is mainly targeting Small Medium Businesses (SMB): cheap, low-end, plug-and-play connectivity. Yes, there are high-end datacenter deployments looking at 10GE iSCSI solution, but very limited, compared to SMB space. -- FCoE can be see as “equivalent” to iSCSI as it runs over Ethernet as well. But, first, FCoE doesn’t require the burden of TCP/IP as iSCSI, resulting in a more agile and less host performance dependent technology. Second, FCoE is meant to run with existing high-end Fibre Channel infrastructure. Today Nexus 5000 is consolidating LAN and SAN traffic from the hosts to split it to Ethernet and guess what? Fibre Channel. This clearly optimizes the access layer utilization. In the future, Nexus 7000 will be able to provide an extra aggregation layer to hand off FCoE traffic to, guess who? MDS 9000 products, i.e. the Fibre Channel SAN. -- Fibre Channel is the storage protocol by definition and it’s here to stay, as well as the MDS 9000 Family. I don’t even wanna spend time on discussing this, it’s really obvious. Just think about the 2.5 million ESCON (yes, not even FICON) ports still out there. How can anyone even think about Fibre Channel disappearing? From a Cisco investment perspective, I believe it’s clear how we will play, just by looking at the history and principles of the company. Cisco has always been technology agnostic and made savvy investments on the most promising ones. We invest in iSCSI (well, transparently -- we don’t even know if that Ethernet port is used for storage or not, it’s just an Ethernet port!), we certainly invest in FCoE and we obviously continue to invest in Fibre Channel and the MDS 9000 Family, where we own 50% of the high end market. Customers want to decide which technology better fits their needs, they want options. And we are here to give those options. All of them, because this way, whatever technology customers pick, Cisco always win: an iSCSI installation, an FCoE installation or a Fibre Channel installation. Hopefully this simplified picture will help get a better understanding of what the future of Fibre Channel and storage networking will be.