3dmentat - Fotolia
If after recent Fibre Channel revenue announcements you thought plucky little Brocade was going to be crushed, flattened under Cisco's Ethernet steamroller, with its Fibre Channel base swept away in a rush to FCoE, think again. Brocade is sitting pretty, its Fibre Channel base is secure, and its Ethernet vision is strong and healthy.
How does this compute? Well, Fibre Channel users will not convert to Fibre Channel over Ethernet en masse; it isn't going to be simple.
What will you need to deploy FCoE? First, there is the little matter of running new Ethernet cables between servers, switches and storage arrays. Then there is also the new Ethernet standard, Converged Enhanced Ethernet (CEE), the multi-hop capability needed for end-to-end Fibre Channel over Ethernet server-switch-array communications. And there's all the clever SAN fabric stuff done in Fibre Channel switches and directors that will need to be done in FCoE-capable switches.
None of this is straightforward, and if the Fibre Channel fabric per-port costs can be kept within a reasonable distance of CEE per-port costs -- with speeds the same or faster -- than Fibre Channel fabric customers won't feel compelled to move.
Fibre Channel customers do not face a strong necessity to move to Fibre Channel over Ethernet. At the same time, we all think the long-term answer to the what-to-do-about-Fibre-Channel question is FCoE, so no new supplier is going to enter the Fibre Channel switch, director and fabric market.
Three suppliers -- Brocade, Cisco and QLogic -- have the market to themselves. Two of them see their future outside Fibre Channel because that's where their future growth is. Cisco wants to kill Fibre Channel with Fibre Channel over Ethernet. Ditto for QLogic, plus it has its InfiniBand business. Neither of them will invest a lot of money to ensure that their Fibre Channel products are better than or equivalent to their Fibre Channel over Ethernet and InfiniBand products.
Brocade's situation is different. It has a solid and well-entrenched position in Fibre Channel, having been the main driver and beneficiary of Fibre Channel supplier consolidation. It has also enunciated its Brocade One converged and virtualised data centre vision, of which Fibre Channel over Ethernet is a strong component, though only a part. The vision is about the whole data centre and the network to connected users with desktop and mobile devices.
Brocade Fibre Channel customers know they will be able to move to Fibre Channel over Ethernet within the Brocade product set and that Brocade has a vision about data centre networking and virtualisation like Cisco. It has a good integration story with VMware, which it's going to spread over to Microsoft Hyper-V and Citrix XenServer. It will front-end Fibre Channel fabrics with FCoE-capable Ethernet and help its customers integrate iSCSI and NAS storage with Fibre Channel devices.
So what can Cisco or any other Fibre Channel over Ethernet supplier offer users that they can't get from Brocade? It's not like Brocade is hindering any move from Fibre Channel; it's helping it whilst keeping Fibre Channel hot.
They know that Brocade will enhance Fibre Channel; it has 16 Gbps Fibre Channel gear coming within a year, and that will double the current 8 Gbps Fibre Channel. QLogic and Cisco may well follow suit, but are not likely to be as enthusiastic about the move as Brocade.
Cisco relies on having higher margins on its networking gear than does Brocade. If Cisco wants to compete hard against Brocade's Fibre Channel offerings head-to-head it will have to lower prices, and that could hurt its business model more than Brocade's.
It's easy to see Cisco effectively walking away from Fibre Channel, keeping a token product presence, selling Fibre Channel over Ethernet against Fibre Channel and saying to Fibre Channel customers, "You have to move to FCoE." Meanwhile, Brocade can say, "Move to FCoE when you want. In the meantime, here's this faster, better, easier-to-use Fibre Channel product. Oh, and if you do want to move to FCoE, our vision and product set is cast in similarly strong metal to Cisco's."
Nobody is positioned to sell into Brocade's Fibre Channel base effectively at all, it being Brocade's to harvest or lose for the next five years or more. What a great position to be in! Brocade faces neutralised competitors and has a set of server OEMs unwilling to commit wholesale to Cisco or HP (ProCurve) networking because both of them represent strong competition on the server front and they need differentiation.
Brocade's storage partners will be willing to keep good relations because they dislike and fear the EMC part of the Cisco/VMware/EMC monster, and they will be cautious about HP's StorageWorks being integrated with ProLiant Servers and ProCurve networking.
What this means is that Brocade can tap into a set of server and storage supplier partners that are uneasy about Cisco and HP, and that want a measure of independence. Add to this Brocade's Fibre Channel market becoming a secure fief and Brocade One being a very partner-friendly converged and virtualised data centre vision, and it becomes apparent that Brocade has played its strategic cards very well, giving it a set of tactical advantages.
Don't pity Brocade, or feel sympathy for its predicament. Admire it instead. Michael Klayko's company has been very smart indeed.
Chris Mellor is storage editor at The Register.