Kumar Malavalli, Cofounder, Brocade Communications

"Cisco might be a big name in routers, but it's not in storage. Data center people want a complete solution, not just networking."

August 20, 2002

8 Min Read
Network Computing logo

Kumar Malavalli is the cofounder of Brocade Communications Systems Inc. (Nasdaq: BRCD) and advisor to Greg Reyes, Brocade's chairman and CEO. He is Reyes' eyes and ears (and nose and throat) for technology. In a nutshell, he has overall responsibility for shaping Brocade's technology direction and product strategy.

Malavalli is also an angel investor in Alpine Technologies, Cloverleaf Communications, InterSAN Inc., Karthika (acquired by Kasten Chase), LightSand Communications Corp., and SV Systems -- interestingly, all storage networking startups.

Why did he invest in these companies in particular, you ask? Is there a link with Brocade? All will be revealed in good time, patient reader. First a little background on our latest Byte and Switch interviewee.

Prior to Brocade, Malavalli was Hewlett-Packard Co.'s (NYSE: HPQ) senior Fibre Channel architect. A key contributor to the industry team that originated the Fibre Channel architecture, he guided the technology through the industry standards committees and holds several U.S. patents in Fibre Channel.

In 1995, a Silicon Valley venture capitalist, Seth Neiman of Crosspoint Venture Partners, contacted Malavalli after hearing of his work in the area of FC. Neiman, with Malavalli's help, went on to launch Brocade in August 1995.The rest, as they say, is history. Brocade underwent a successful IPO in 1999 and has emerged the reigning champion of Fibre Channel SANs, with well over 50 percent market share today. But the picture is changing. New storage networking technologies are evolving, backed by technology giants ten times the size of Brocade, while innovation among startups in this market is astounding. Can the original FC company hang on to its leadership as storage networking moves into a new generation?

Kumar Malavalli talks to Byte and Switch on this and other hot issues in storage networking. Click on the links below, or read the interview sequentially.

Byte and Switch: Tell us about your investments and why you have focused on these companies in particular.

Malavalli: They make up my SAN ecosystem picture. Brocade cannot do it all alone; these companies fill in the holes and take them to a higher notch.

[Ed. note: Most of these companies are still in stealth mode, but to run through them briefly: LightSand offers a Fibre-Channel-over-Sonet gateway; SV Systems is working on automated disaster recovery for mid-tier systems; InterSAN provides automated discovery, monitoring, provisioning and control of storage network devices from an application perspective; Karthika is building a storage security appliance; Alpine is building an "intelligent affordable storage subsystem," Malavalli says; and finally, Cloverleaf is working on application-aware storage management.]Byte and Switch: There seems to be some overlap between InterSAN and Cloverleaf in the storage management software arena, no?

Malavalli: InterSAN does path provisioning, mainly, whereas Cloverleaf is focused on application-aware resource management. There might be a slight overlap, but as long as there's no duplication, it's OK. I host a SAN systems roundtable every two months for all my startup companies and invite the engineers and marketing people from each to talk through what they are doing to make sure they are all working together.

Byte and Switch: Is that something they look forward to, do you think?

Malavalli: Well, I hope so... We meet in the Bay Area, which is closest to where they are all located. The idea of the roundtable is to create a virtual company with individual ecosystem members still maintaining independence in their special area of expertise and product development. The companies can share common tools, which are outside of their intellectual property, and work towards providing an optimum solution.

As a group, we hope to eliminate barriers of adoption by end users, who have a history of looking to large companies for solutions. We also want to eliminate fear and doubt to customers in considering incorporating products from startup companies into their data centers. We are attempting a very innovative technique to make several small companies a virtual large-scale company.Byte and Switch: Where does Brocade fit in? Is one of these companies going to be the next Brocade?

Malavalli: Another Brocade? That's hard to envisage.

Byte and Switch: Are you are lining these companies up as potential partners or acquisition targets for Brocade?

Malavalli: I invest based on the holes in Brocade, but I do not want to push the company into anything. They must take a fresh look and see if it's worth holding hands. The beauty is, I am taking all the risk.

[Ed. note: This is an interesting definition of risk. Keep in mind that Malavalli is working for Brocade. Should his investments not work out, he may be able to write them off -- at least, part of them -- through Brocade. Then again, should Brocade decide to acquire one of these companies, Malavalli wins twice over: as an investor in the startup and as a significant shareholder of Brocade. Put simply, there is little to no risk with these investments. Malavalli declined to disclose the total amount he has invested.]Byte and Switch: Would you say Brocade is an innovative company?

Malavalli: Yes, of course.

Byte and Switch: Why?

Malavalli: Creating all the ingredients to manage an intelligent distributed fabric, which is what Brocade has in Fabric OS, takes constant innovation, and this is going on at the company all the time.

Byte and Switch: Brocade doesn't like to talk about new technologies that might threaten its installed base -- iSCSI is a good example. During a Salomon Smith Barney technology conference last September, Brocade chairman and CEO Greg Reyes said:

  • "Our installed base will be so massive by the time IP or InfiniBand becomes a reality, that no company will be in a position to compete with us on infrastructure or applications." (See Brocade CEO in Tick-Top Form.)

It sounds like Brocade has a classic case of Innovator's Dilemma, the phenomenon coined by Clayton Christensen. At the heart of his book is the issue of how a successful company with established products keeps from being pushed aside by newer, cheaper products that over time, get better and become a serious threat.

It should be on Brocade's reading list, don't you think?

Malavalli: Of course, but listen: Greg and I used to discuss iSCSI all the time and I was telling him it's a nice idea but it's not going to meet the total end requirements of customers in the data center.

Byte and Switch: Can you elaborate on that?

Malavalli: Sure. When the Fibre Channel standard was developed, it was designed to combine network and channel technologies into one hybrid technology that supported both master-slave and peer-to-peer paradigms. This was necessary to provide networking aspects to storage and to provide an I/O channel-networking infrastructure. This did not exist in the IP world, which supported only peer-to-peer connectivity required in the conventional networking environment.You are right, in that the introduction of iSCSI has changed that picture. However, Fibre Channel has added characteristics -- with hardware assistance to maintain high performance and throughput -- necessary for the master-slave paradigm. These are:

  • Support for large block data transfers. Fibre Channel "sequences" equivalent to IP "packets" could be very large -- multiples of megabytes.

  • Keeping track of large number of I/O transactions independently with mechanisms to provide independent interlocking between command and status for each of these I/O transactions. Fibre Channel "exchanges," equivalent to TCP sockets, support multiple thousands of concurrent I/O streams.

  • Support for segmentation and reassembly of large blocks of data at a lower level, without needing the intervention of an upper-level protocol, to help transfer the data directly to user space without additional copies thus providing high performance.

Byte and Switch: Are you saying that the Fibre Channel players are adding features to FC to keep up with iSCSI?

Malavalli: I am not a Fibre Channel bigot. I am not married to the technology; I just want to do what's practical and useful to the industry. IP is not a bad thing, but it is meant for certain types of networks. I do not believe one technology will do everything in the world. Fibre Channel offers the highest performance and throughput for storage networking and is constantly being improved -- I don't see this changing.

Byte and Switch: What about Cisco Systems Inc. (Nasdaq: CSCO) gunning down on the company and looking to leverage its massive installed base of customers? We've seen what happened when Brocade and Cisco attempted to work together. It was like trying to get Tom and Jerry to sit down and be nice to each other. It didn't last for more than two minutes. Cisco is preparing to do battle; do you think Brocade is ready? (See Cisco to Christen Andiamo Family and Cisco and Brocade: This Means War.)Malavalli: Of course. But Cisco is not a storage company. It is coming into the market at a place where Brocade was, and by the time it gets to general availability, we will be in the applications space offering higher-level services. Cisco might be a big name in routers, but it's not in storage. Data center people want a complete solution, not just networking.

Byte and Switch: But you'd agree that Cisco is a threat?

Malavalli: It keeps us honest.

Byte and Switch: Changing the subject, what do you do for fun?

Malavalli: It's tougher now. I am working most weekends just to keep track of everything, so it's hard to find the time. I play golf, and I do philanthropic work in education and e-learning.Byte and Switch: What about Pulsar Ventures? You're a partner there, too, right?

Malavalli: Yes, that's right, but only in a small way. It takes up about 15 percent of my time, but it's something I am interested in. Pulsar is an incubator for nanotech and biotech companies. We take the technology developed in research labs and universities and turn this into a business plan. It takes a lot of due diligence, so we only invest in one or two a year with a small fund.

Byte and Switch: Have you had any success?

Malavalli: Our latest company is a project that came out of Penn State University

Stay informed! Sign up to get expert advice and insight delivered direct to your inbox

You May Also Like

More Insights