Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : All About Sun Microsystems -- Ignore unavailable to you. Want to Upgrade?

To: DownSouth who wrote (32894)	6/18/2000 7:43:00 PM
From: rudedog	Read Replies (1) \| Respond to of 64865

I've been following NTAP for a while - first became aware of them because they use Alpha processors in their machines. NTAP's revenues are about 1/10th of CPQ's storage business, and probably less than CPQ's NAS business, although CPQ does not break that out. From a technology viewpoint, there is nothing to set NTAP apart - that stuff is pretty easy, based on Linux, SCO Unix, Novell, even embedded NT. Likewise from a storage management perspective, by definition NAS uses the same infrastructure that a file and print server would use. Maybe I am overlooking something... if so, maybe you could give me a clue. SUNW has a great name in the business, has just announced some pretty competitive offerings, and has a great management story. HP could also be pretty credible. Intel has made noises about enabling a pre-packaged NAS appliance, or at least a board set well suited to doing that job. I just don't see why NTAP would have a lot of running room.

To: DownSouth who wrote (32894)	6/18/2000 8:30:00 PM
From: rudedog	Read Replies (2) \| Respond to of 64865

DownSouth - a few more comments on NAS and NTAP - which are IMO not OT at the moment given SUNW's recent announcement. First a few credentials - I don't know if you know my background, but my undergraduate degree was in microprocessor design, and my graduate work was in control theory. I actually did some microprocessor work for TI in the early 80s but quickly moved to design of specialized disk controller subsystems. I also did a lot of work in the software stacks that interact with subsystems, both at the OS and application layer. I have stayed involved with a lot of the basic research in this area and routinely talk with senior research people at IBM and CPQ about their thinking. Despite the positioning taken in the piece you refer me to, the patent work done by Hitz and Lau was interesting but hardly disruptive. The Unix file system is easy pickings and many different (and patented) techniques exist to add robustness, improve performance, and virtualize access. Likewise NT offers a lot of optimization potential. Novell's systems are pretty efficient - they have always flown close to the machine - but even there, the underlying hardware and system software can add capability. As early as 1993, CPQ's "smart array" controllers were virtualizing the underlying storage to provide higher performance and more robust error management. Their systems include advanced failure prediction (which allows fully functioning disks to be identified as likely to fail, and provides warranty replacement), on-line volume expansion, and abstracts the underlying storage layout with no performance penalty. Those systems have been on the market with substantially those features since 1993, although steady improvement in capability has been made over the years. CPQ has done most of their advanced engineering in the SAN space rather than NAS. Part of this was "elitism" on the part of the research teams - they felt like the inherent penalties of the NAS architecture were great enough that a better architecture would pay off. They may still be right, but there seems to be a place for NAS too. Especially in cluster file systems, which Digital invented and has successfully marketed for more than 20 years, the SAN approach has benefits in both consistency and performance. Cluster file system devices which support a fully snapshot capable view for up to 96 nodes are available today for NT, Unix and VMS - all at the same time. Since I don't want JC to hit me for being "too long" I will close off - but my sense is that NTAP has some good wrinkles, has executed well, but is a long ways from being a "gorilla" and will face some serious competition in the near term, with little aside from hustle and marketing savvy to maintain their position.