Who could buy Brocade ... and what does it mean for customers?

According to reports in the press in early October, the Fibre Channel specialist Brocade has quietly put itself up for sale. The proposed price is said to be in the range of three billion dollars. The company itself refuses to comment and no potential acquirer has come forward or publicly made an offer.

Brocade has basically two main businesses: Fibre Channel switching and the backbone network switching of Foundry.

The Fibre Channel business

Concerning Brocade’s Fibre Channel business, it is important to take note of several points:

- This market is nearing the end of its life cycle. Five years ago, Fibre Channel was the main SAN technology for connecting disk storage to servers. Today more than 50% of the storage sold does not use this technology. Currently popular solutions (SAS attachments for example) don’t need it.

- Brocade has consolidated nearly all of the Fibre Channel market players through acquisitions of companies such as McData, CNT, etc, accumulating a very diverse installed base of products. The perspectives of evolution for this base are of course real, but limited and costly for Brocade.

- The trend towards virtualisation has been bad for Brocade: the reduction in the numbers of smaller physical servers has reduced the need for their connection with storage, which is done otherwise within larger servers.

- Finally, the architecture of blade servers could well put the final nail in the Fibre Channel coffin: they are cheaper than the Fiber Channel adaptors that they do not need. Specialist challengers such as Blade Network Technologies are well positioned to benefit from this market “game changer”.

For Brocade, what remains of the Fibre Channel business is an installed base of mostly larger SMP servers connected to somewhat complex SAN storage.

Who dominates this stable market? On the SMP server side, IBM together with HP and Sun. On the storage side, IBM again (less convincing) but also (and especially) EMC, Hitachi and HP. It is probably in this list that a potential buyer could be found, probably as a base protection initiative … but without great enthusiasm.

The Foundry business

Foundry, a specialist in large data center switching, was bought – for three billion dollars - by Brocade in July 2008.

The main idea was to move beyond the Fiber Channel segment into high end network switching. With this strategic move, Brocade was positioning itself in direct competition with Cisco which also offers fiber channel (some) and networking (a lot). In this way, Brocade could become a central player for communication between servers and storage in all sorts of combinations.

Unfortunately, effective execution by Brocade was handicapped by a combination of negative factors: a slowly growing market, a somewhat limited innovation capacity and, perhaps most important, the distractions of internal work - arising from Brocade’s different acquisitions - of rationalising various technologies and a complex installed base. In this context, Brocade failed to « catch the wave » to surf on a truly major technology trend – the move to virtualisation.

Of course, virtualisation has had both positive and negative implications in this domain for Brocade:

- the upside: the consolidation of older switches onto the larger equipment that Foundry provides (but also Cisco and Juniper)
- the downside: "real communication" has often been replaced by "virtual communication" thanks to the hypervisor. Here, Cisco and EMC/Vmware have done well, while Brocade has been competitive only on a part of the solution.

Since Brocade is essentially present only in the "real" world, it is not pertinent on the growing part of the server market that has become virtualised. In the real server space, it is in competition with players such as Cisco, which has an image of greater solidity in a critical domain where solidity is a key decision factor for customers.

There are potential buyers but none appears to be an obvious fit for Brocade:

- IBM could be interested in order to avoid the disappearance of a competitor to Cisco, in addition to the interesting Ficon (zSeries) support within Brocade;
- HP could be somewhat interested for its high-end servers, but only marginally;
- Sun has other problems and Oracle too. It is probably not the moment to take on Brocade’ difficulties;
- EMC could consider such an acquisition with interest, but this would raise problems with Cisco. The timing is not right. They owned Mcdata in the past and dropped it...
- Hitachi is somewhat marginal in this particular game and its storage offerings are multi connection. Why make life more complicated?

Whoever the future potential buyer might be, a price of three billion dollars - while somewhat lower than the current market cap – is almost surely too much for too little in terms of business fundamentals.

In addition, the fact that all the potential buyers are Brocade’s OEM customers - in competition with one another - means that any such acquisition could lead to a significant loss of business for Brocade.

While we certainly do not believe that Brocade is in mortal danger, Brocade customers would be well advised to undertake a review of their data center networking including:
- take a fresh look at switching in the data center, taking into account the probable evolution of servers and storage over the next 18 months;
- determine where Fibre Channel is used and should stay in place over the next 18 months;
- study the potential of IP based solutions (iSCSI) and other similar technologies;
- examine the alternatives on these points (for example QLogic, Cisco, Blade Network Technologies);

Given the rapid evolution of data center infrastructure, such a review might usefully be enlarged to include additional actions such as:
- study and plan substitution technologies for storage and servers – in the short term, in the 18 months to come and beyond;
- consider the carbon footprint and the energy aspects in the data center;
- simulate what your network is looking like after you virtualise 25% of your x-86 servers, 50%, 75%.

Overall, customers should understand that while Fibre Channel will continue to be necessary in many data centers for some time, it should now be regarded as a tactical - rather than a long term -investment decision.