It was a small group, only four, so he asked us all to introduce ourselves before he started.
And when I introduced myself, there was a slight problem. This sort of session only works if participants speak their minds freely. Journalists are usually not allowed to sit in on this kind of gathering. However, once I agreed to identify participants only as A, B, and C and be extra-careful not to include any clues to their companies' identities in this story, I was allowed to stay.
I would like to thank those company representatives -- and Mr. Fichera -- for allowing me to participate. My carefully edited account of the proceedings follows.
Sun is loved less than other vendors
Companies A and B are in the communications industry. Company C is in the computer service business; it resells and maintains Sun (and other vendors') hardware and software for a variety of clients. None of these companies is exceptional. They are old-line Unix shops that have been using Solaris for years and like it, but do not expect to keep buying from Sun as heavily as they have in the past. All three say reliability, not price, is their primary concern.
Fichera displayed a chart showing that 59% of survey respondents said they were "Happy overall" with their hardware vendors' product quality, but only 44% of Sun users were this full of joy. (IBM, HP, Dell, and several other hardware companies were also included in the survey). It was sad to see Sun fall so far below the average here, and even sadder that A, B, and C all agreed that this was a plausible result; that they were not as happy with Sun as they were with some of their other vendors.
Fichera then asked if the 2000 - 2001 Sun cache memory problems (and Sun's initial denials that the problems existed) were still a factor in anyone's perceptions of Sun's product quality. B said the defect, compounded by Sun's poor handling of the problem, was a major reason his company started looking for alternate sources of SPARC servers -- and ended up buying a number of Fujitsu products. He also said his company has used references to past Sun hardware problems combined with the threat of an alternate source of supply "as a hammer" to win price concessions from Sun.
Company B has had no problems running Solaris on Fujitsu servers. The only "software problem" they've had really didn't have anything to do with the software itself: IBM refused to certify WebSphere on Fujitsu servers, even though Company B found that it ran as well as it had on Sun hardware. (IBM was, however, eager to come in and help port any or all of B's applications to AIX at little or no cost.)
All three agreed that despite some famous quality problems (and terrible customer service when they asked Sun to fix them), their entire Solaris hardware plus OS plus applications stacks were giving them 99.7% uptimes, and if you removed applications and network attachments from the equation, and looked only at the hardware and Solaris itself, they were getting 99.99% or better. One said, "We're confident in the operating system. It's our preferred operating system. But we have reservations about Sun's longevity."
Worries about Sun's survival
Nobody wants to be stuck with a server room full of orphaned boxes, and companies A and B noted that doubts about Sun's future are a major reason they're considering alternatives. Company C, as a service firm, installs whatever its customers request, so while it has a strong Unix track record and an internal preference for Solaris, if Sun goes away it will simply switch to another vendor -- and charge customers to replace Sun boxes (and possibly Solaris) with whatever else they choose.
Company B is experimenting with Linux as a possible long-term Solaris replacement, but right now is comfortable with Linux only in low-level applications. A said his CIO was not interested in Linux at this point, that he's "not fully on board the Linux train," so if they move from Solaris it will be to AIX or HP-UX, and that they are currently using Linux only in some very limited circumstances. Windows server products are not under consideration at either A, B or C. These companies are all familiar with Unix, like Unix, and plan to stick with Unix. They see no major performance differences between the three big-vendor flavors, but point out that more ISVs (independent software vendors) work with Solaris than with other Unixes, and that more sysadmins have experience with Solaris than with the others.
Fichera asked again about Linux, this time pointing out the huge price/performance difference between Linux on AMD and Solaris on SPARC. The answer he got was that while the price difference was great, not enough ISVs were certifying their software on Linux to make the jump attractive quite yet.
Next, Fichera asked if the Linux temptation would be reduced if Sun lowered Solaris/SPARC prices enough to compete with Linux/AMD. The Company A person said, "They probably can't do that." He believes Sun has slashed prices on its core product line as much as it can afford -- and then some.
Sun's published price lists have become meaningless
Forrester's survey showed 48% of the respondents were satisfied with their vendors' prices. Only 33% of Sun's customers said they were satisfied. And this is despite all three men in the room saying they routinely ask for -- and get -- discounts in the 50% range on every Sun hardware purchase. One said he hears that some of Sun's largest customers are sometimes getting 60% discounts.
None of them thought Sun could afford to discount any further and ever make a profit.
(These gentlemen and their companies may have doubts about Sun's survival under its current price structure, but that doesn't keep them from taking advantage of Sun's vulnerability by timing as many purchases as possible so they come at the end of Sun's business quarters, when Sun reps are likely to give even deeper discounts than usual in order to "make their numbers.")
Fichera wanted to know if anyone felt Sun was trying to make up for hardware pricing problems by raising software prices. B said yes, they were trying to "squeeze on software" but that it was not a very successful attempt. A said Sun was "constantly knocking on the door, trying to sell professional services," but wasn't getting anywhere with his company, or with other large companies as far as he knew. Fichera said that from what he had heard, Sun was having at least some success selling additional software and services to smaller companies, and got agreement from A, B, and C that this was possible, even likely. (A company with only one sysadmin can be very receptive to a service contract pitch if that sysadmin quits.)
The last question from Fichera was whether people felt Sun made a real attempt to understand their businesses. A said, "They understand our business because we forced them to." B said, "No. Our relationship with them has deteriorated to the point where they don't make the effort." C laughed.
A, B, and C all said they've seen changes in the behavior of Sun's sales force in the last year or so. All said that in their experience local Sun sales and support people were generally decent and professional, but that when they had to deal with people at Sun HQ things were not so good. The phrase "like the Three Stooges" was mentioned during a description of one service escalation episode, and brought "I know what you mean" mutters from the others in the room.
But possibly the most telling (and saddest) anecdote regarding Sun's recent sales and service behavior was one of the session attendees' recollection that late last year their Sun rep called on them and announced, "We have a new culture. We have to compete now."
Why, we all wondered, did it take Sun so long to figure this out?
Nobody had an answer, and it was time for the session to end, so we left it at that and moved on.