Slashdot Mirror
Should Sun Just Fold Now?
KE1LR writes "The Silicon Insider at ABCnews.com is taking the position that Sun Microsystems, creator of the SPARC architecutre and, oh yeah, Java, should just give up and close shop instead of continuing to wither. I agree that Sun would have to have to do something dramatic to avoid what is looking more and more like an inevitability at this point, but what could stop this slide toward the same fate as DEC? Might they have anything in the works that could save them? What could it be?"
yeah fold NOW
$ strings FTP.EXE | grep Copyright
@(#) Copyright (c) 1983 The Regents of the University of California.
Just a bit of info:
Sun's stock (SUNW) is now hovering at about 4.00 (down slightly today).
Here's SUNW over the past 5 years
Casual Games/Downloads
Sun insists that they won't sell Java to IBM. IBM is now quite dependant on Java and have all sorts of ideas for how they would like to change it if they didn't have to constantly butt heads with Sun.
So, okay, fine, IBM can just wait a bit and buy Sun for a reasonable price. That way, Java won't have been released into the public domain and IBM won't have to argue (as much) when they want to change it.
IBM has the most to gain from control over Java -- arguably Microsoft has more but for legal reasons they won't bother even trying to buy Sun -- so they'll be willing to pay the most, so they'll get 'em.
"Those who have never entered upon scientific pursuits know not a tithe of the poetry by which they are surrounded."
Now, I'm going to watch the presentation on Croquet.
There is a reason for including these utilities. Sun prides itself on maintaining backward compatability with pretty much everything written for its OS back to say Solaris 2.6. If they kept changing the utilities, scripts that leverege these utilities would break. This is a HUGE deal for companies that run legacy apps.
Yeah, the Solaris tar does suck, but on my JumpStart server, I ALWAYS include the GNU versions of tar, gzip etc.
If you want the latest-greatest, load the GNU utilities from the Solaris Supplemental CD. Or download them from www.sunfreeware.com. Or compile them yourself. Or install them via RPM. Sheesh.
"The greatest dangers to liberty lurk in insidious encroachment by men of zeal, well-meaning but without understanding."
I'll wager this sort of eye-for-an-eye, zero sum logic is endemic to human thinking altogether.
Human history is chock-full of one culture declaring another ( largely similar ) culture to be at odds with their god/economics/what-have-you and proceeding to at least try to wipe the other out.
Think: crusades. Think: cold-war. Think: carthago delenda est.
The zero-sum is not exactly new. The difference is now we're dealing as much with corporate entities as with foreign cultures.
lorem ipsum, dolor sit amet
Their customer support sucks. I say let Sun evaporite in a wave of Hawking radiation.
Just a back up to the pp.
Apple doesn't scale because you can't put a large number of chips in the same box. x86 isn't as limited, but it's not great either.
As you add processors, there is a diminishing return on inverstment with each one. iow, two uniprocessor boxes will be able to do more work than one dual box, however they cannot operate on the same data at the same time (I know beowulf, etc. give me a break). In some cases, one box with n chips will outperform a box with n+1 chips.
On Sun hardware, this difference is less than on apple and x86 hardware.
Sun's architecture is designed from the ground up to have a bunch of processors in the same box. This is part of the reason that their uni boxen are unimpressive performance wise. Scalability sometimes hurts small scale performance(Think using Oracle/MySQL/PGSQL for a table with 100 rows. Sorting and binary search would be faster).
Have you ever read through Sun's FE Handbook? It's a nightmare. Ever tried to hot-swap hardware inside a production Sun server while it's online and in use? Bah. Give me a room full of Linux PCs any day!
Google spent... oh, roughly $100m in software development getting to the point that they were saving enough money by using the distributed low cost low reliability PCs. That is a huge barrier to entry on such largescale clusters.
And Google is in a business where a little data loss in the searches is not going to seriously harm anyone. So they operate slightly lossy. They admit this pretty explicitly; one of their people, Anurag Acharya, was an invited speaker at the second Evaluating and Architecting System dependabilitY symposium in 2002.
Neither the software investment to make reliable distributed apps nor the lossy data model are acceptable to typical business software. Do you want your bank losing 1-2% of your deposits, or having a consistency check error balancing your account at the end of the month? How about Amazon randomly deleting or inserting a few things from your orders...
And even where there is off the shelf distributed software like Oracle RAC, it's such a management and performance hit that people typically go back to buying larger single system image servers after testing it out... ask Oracle what percentage of their sales are RAC versus straight Oracle 9 some time.
There are applications... web farms spring to mind... where the Google model is a natural fit for the problem set. Strangely, that particular answer was well known five years ago, because people are not stupid.
Until every major business application is naturally and easily distributable larger servers will continue to sell. The software is just plain not there yet. Things are trending that direction; in ten years, the current model is in real serious trouble. Maybe sooner. But now? Don't believe dumb hype.
Ok, so let's compare. Let's compare a Sun Fire V440 and a HP DL580 G2. Let's assume each is equipped with 4 top end CPU's, 8GB memory, dual Gigabit NIC's, 2x36GB disks, and a DVD-ROM drive on each -- sounds like a fairly standard server configuration to me.
Price
The V440 is more than 50% less!!!!!!!!! Ok, let's go to performance. Going to use the SPEC CPU2000 info for the DL580 G2 3.0GHz Xeons and going to use the Sun Fire V250 config mutltiple by 1.8 (since Sun has not yet releaed info on the 4-way V440 with the same 1.28GHz US IIIi CPU's tha the V250 has). (Listing below represents Cint2000/Cfp2000/Cint2000 rate/Cfp2000 rate).
Performance
Hmmmmm....two things jump out at me here -- the UltraSPARC IIIi is lousy at integer math, while the Xeon is lousy at floating point math. Either way, the 3.0GHz Xeon, which represents a clock speed difference of 234% greater than the US IIIi, only performs better than it by 28.7%. Increasing the CPU to 1.7GHz or going to US IV CPU's as Sun plans to do with the upcoming V490 will close the gap.
So overall, for 109.6% of the price of a V440, you're only getting 28.7% of the performance. Umm....what was your original point?
Rule #1 -- Politics always trumps technology.