Did he get tired of waiting for his stock options to see daylight? Perhaps reporting to a guy with a ponytail didn’t appeal – HP guys can be funny that way. Or maybe five years in the Mos Eisley of storage finally broke him. Something, clearly.
An oddity: Sun says that David Yen is taking over Sun’s Storage Group and that Mark Canepa “former head of Sun’s Data Management Group” is leaving Sun after 10 years. You have to rummage around on the website to discover that the Storage Group and the Data Management Group are one and the same. The slyest name change I’ve seen in years.
Could it be that Schwartz, a software guy, didn’t like the DMG name, since it sounds like software? And that he doesn’t like Yen, probably the premier hardware guy at Sun after billionaire co-founder Andy Bechtolsheim, who gets to do whatever he likes.
The loser in all this is Yen, a smart and capable guy. Picking up the pieces of the long, slow train wreck that is and will be the StorageTek acquisition won’t be fun. He does have Project Honeycomb, which has the potential to start reshaping the entire industry. Let’s hope he sees that and drives it aggressively.
If Yen can breathe some life into a “dead horse”, more power to him. His track record has been pretty good in the hardware area. Sometimes late to very late but the “Right Stuff” delivered. I got pleasantly fooled a couple of times expecting the usual Sun “Misses”.
Sun Storage, IMHO, has been an albatross and an anchor on Sun Microsystems integrated solutions. They need good storage. Gosh knows they have bought some of the best Storage software companies I ever saw.
Hopefully they can do a better job of integrating the reasons they bought these companies than they have. I certainly hope they can do a better job than EMC which has absolutely destroyed some of the finest products I ever saw.
Bottom line? I see Yen as the only hope they have from the inside. They show no desire to go outside.
I once ended a Sun presentation with the question, “You have all the products with the right feature/functions and the right integration. Why are you not leading the market?”
They were offended, as was my boss, and ended the presentation without answering. My boss ended my career saying it was bad form to.attack people who were invited to give a presentation. That’s when I knew I never had a career, only a job.
From my point of view Storage is about at the level of evolution the abacus is for calculators.
David Yen is a solid guy. Sadly, he is being ushered into a group that has defeated the best efforts of several different but capable managers. And Canepa leaving just a few months after the STK acquisition — not a good sign. STK is the past. Honeycomb *might* be the future. I wish David the best of luck.
Let’s hope that David Yen uses the Wayne Gretzky method of, “Skate to where the puck is going, not where it is”. So where is Storage going? Nowhere fast! Incremental change has been the best revenue producer for Storage vendors for a long time and they see no reason to change it!
Let’s take STK. There are some real bright spots in STK. The rest is “the past”. On a good day the “20%BrightNew/80%DeadPast” Rule applies. On a bad day the “2/98” Rule applies. I hoped at first the buy was made to get those “BrightNew” spots, integrate them and be a market leader. Not so. The buy was made to get the incremental business to boost the bottom line.
Maybe, rather than attempt to integrate the “2-20%BrightNew” into Sun’s “80%DeadPast”, David Yen should just whip out his Star Wars “light sabre” and slash the “Gordian Knot” of “DeadPast” Storage. There goes incremental change.
Most of the Storage startups of the last few years, who proclaimed loudly they were going to rewrite the future of Storage, were just rearranging the deck chairs on the Titanic or Hindenburg. Take your pick.
Now we have the incremental solution of “tiered” Storage.
For example, if I need “99999 (five nines)” of Information High Availability (IHA) in my Primary Storage, and maybe I need “99999 (five nines)” in my Primary Mirror, how many “nines” of IHA do I need for the third and fourth mirrors at the geographically dispersed Disaster Recovery sites? Could cheaper Storage be used for these mirrors? Could cheaper Storage be used for the Primary Mirror if it was redundant enough? Like clustered? How cheap would it have to be?
Then there is Content. What a national disgrace. Why can’t the Storage box figure out the Content of its contents on the fly? All the Storage box needs is a little help from the application. A few little objects attached to the packet.
There are only two rules for Content.
1) Which of your Content generates 80% of your revenue?
2) Which of your Content, should you lose it, puts you out of business?
And tape. Why do I have to use tape?
Or disk? Disk is the best solution, for the Enterprise shops who care. And for shops that have very short RTOs. So short an RTO that they cannot afford to ever go down. Not even for a nanosecond.
What about RAIDVD? Oh Gee! The media only has a two year archive life? I bet if DVDs had the same archive life as tape they would still be cheaper to buy, burn, and send to Iron Mountain. The DVD drives certainly would be.
[Note: RAIDVD does not use carousels.]