Playing the API card

Brocade Communications Systems and its champions have promoted the debut of the company's new switch API to justify Wall Street's renewed optimism. The argument is that if the company's products can better integrate with other storage applications and management software, Brocade's popularity will increase. Suppose it were so. Still, can an API ever be so compelling or powerful that it justifies an increase in a stock's valuation?

An API is certainly a worthwhile thing to develop. It extends the potential usefulness of products by allowing them to integrate with future storage network management applications. However, that's a long way from saying that a switch API is the key to the future of storage management.

Switches forward network traffic. They do not store data. Networking in storage is an important development, but it is not the driving element in storage networking--storage is.

The management of storage will be independent of anything Brocade does in the switch. Brocade switches are part of the storage path, but they do not control storage network traffic--host systems and storage subsystems do.

There is no doubt that placing SAN (storage area network) management functions in the network is a good idea. However, it is not clear that putting management functions in the data path, where they can impact and interfere with input/output traffic, is an intelligent thing to do.

Networking in storage is an important development, but it is not the driving element in storage networking--storage is.

Looking past the details of the SAN industry, the value of intelligence in a storage network--or any network--is questionable. The ideas of David Isenberg, who initially established the theory that stupid, fast networks will beat out intelligent networks, appear to be born out during the last two years. One of his most recent thought pieces, "The Paradox of the Best Network," restates his thinking in a way that can be quickly assimilated and applied to storage networks.

There is no reason that Brocade or any other network storage company should expect the situation to change. In fact, the value of intelligence for storage input/output is even more questionable, and the value of switch intelligence, tied to an API, is dubious.

Even if network intelligence in SANs should turn out to be valuable, the Brocade API does not put Brocade in the driver's seat for controlling the SAN--it instead allows Brocade switches to be controlled by some other type of network storage management product from companies such as Veritas Software, Computer Associates, IBM, Hewlett-Packard, EMC and others. The ability to be controlled by network management is hardly earthshaking news, but it is getting as much play as it possibly can for Brocade now.

In fact, the value of intelligence for storage input/output is even more questionable, and the value of switch intelligence, tied to an API, is dubious.

The key to success for Brocade, as it is with any other company, is whether or not they will be able to execute on the fundamental value propositions. Core value is not determined by creating an API that supports hourly traffic reports. All switch vendors will have this type of function or they will be out of business. Imagine introducing a TCP/IP networking product that did not allow network management control. It would be laughed off the face of the earth.

So the Fibre Channel companies are a little behind and are trying to catch up with the rest of the networking industry, which is in serious reverse. Fibre Channel is hardly an industry to get overly excited about. Obviously, Brocade does not view the world this way and continues to look for ways to generate excitement. If your last big ploy to get attention was the underwhelming 12000, maybe this new API can help obscure the doubts about this looming letdown.

The lack of fundamental strength in the Fibre Channel industry is amplified by the selection of an API as the most exciting reason to buy the stock of the industry leader. That should speak volumes.