What are your impressions of McData's recent purchase of Nishan and Sanera? What does this mean for customers?
It was a very smart move by McDATA. The acquisition of Sanera gives McDATA a much higher top-end that scales to 1024 ports architectually with full bi-sectional bandwidth (true non-blocking). It has functionality (dynamic partitioning) that competes effectively with Cisco's VSAN. It's a multi-protocol director (FC, GigE, iSCSI, & FICON) with configurable ports. This product gives McDATA the functionality highground and maintains their leadership in the SAN director market space.
The Nishan acquistion is both tactical and strategic. It provides McDATA a SAN WAN gateway functionality. More importantly it provides SAN routing between SAN islands locally or remotely. This is important for those storage targets that need to be accessed from multiple SANs, VSANs, or partitions.
An example of this are automated tape libraries. Nishan solves this issue while preventing unwanted traffic or disruptions from moving between the SAN islands. McDATA customers get increased funtionality and Nishan customers get increased support, increased product flexibility, increased visibility, increased storage vendor support, and a longterm product roadmap.
Looks like a "win-win-win" to me. In the past you have been critical of intelligence in the switch. Do you think users will have to settle for intelligence in the switch, or will there be vendors that don't do this?
Let me clarify that question. I have been critical of the value proposition for intelligence in the switch, not the concept. If the value proposition or ROI to the end user makes sense, I am all for it. It is when that value proposition cannot be quantified, I am less than enthusiastic.
SAN switches are moving in two directions. The first is commoditization or much lower pricing that will meet the needs of the majority of the market. The second is increased functionality or intelligence which will cost considerably more. The key for both is the perceived and actual value received by the customer. Does Brocade need a 256-port switch to compete? Basically, can a user get more from a "McData, CNT-type" if they buy director and edge switches from one vendor?
The 256-port director market is a small part of the market. McDATA is the director leader and the 256-port director is a necessity to maintain that leadership. Brocade is the leader in the mid-tier switch market. Having a 256-port director would be nice and allow them to compete more effectively against McDATA, Cisco, and even CNT/INRANGE. The question remains do they need it to compete? For the FICON space yes, all others no. With the Nishan purchase, McData now has a IP switch. Brocade has said their roadmap has an IP switch debuting in 2004 (I guess one can argue the Rhapsody equipment is IP-ready). Do you think this will accelerate Brocade's road map? Or, has this market not matured enough yet to make that big of an impact?
It likely will have zero impact on the timing of Brocade's roadmap. The IP SAN market is nascent and nowhere near maturity at this time. CNT buys Inrange, McData buys Sanera, Brocade buys...?
They already bought Rhapsody and are still digesting them. Another purchase would make sense if it has the right value proposition to Brocade and their customers. Any guesses as to who that could or would be, is pure speculation. Can the QLogics, MaXXans, Broadcoms of the world survive with McData, CNT and Brocade providing a full range of products?
The market is not a "winner-take-all" scenario. Each of these vendors have very different strategies with various degrees of success.
QLogic is leading the charge in reducing the cost of SAN switching. They are the leaders in SAN switch ASIC designs. They measureably increased their design wins with both embedded and box switches, and increased the number of OEMS, channels, and SAN switch market share -- albeit, with little fanfare. And these gains are large enough that SUN is no longer their largest switch OEM.
MaXXan is the only director class vendor that is providing integrated and embedded open storage applications in a single image with a single management interface. Their approach (called SANe or storage application network engine) is to provide optimized storage applications tightly integrated with the SAN as a complete solution to the end user. Their focus is on the DR aspect of the storage business and they appear to be striking a nerve with the end users while gaining market traction. The jury is still out on Broadcom. They have been actively going after the embedded space with new designs. They are somewhat late to this space and have numerous challenges to overcome. Broadcom is a savy component supplier. I for one would never count them out.
Brocade and CNT also have significant challenges in front of them. What is the most useful emerging switching technology you've seen in the past few months?
It depends on what you mean by "useful". I define useful as high value proposition to the user. And "most" is a relative term. I have found 4 very useful technologies in the past few months: