The curious case of Sun's hardware biz Today is the day that shareholders of Sun Microsystems will gather to decide the fate of the company, either approving or disapproving of Oracle's $5.6bn takeover of the server and operating system maker. Considering Sun's options - which are basically none if the Oracle deal isn't approved - the fate of the company is not …
I disagree with the statement that SPARC64 barely keeps up with Itanium... That's really quite laughable as Itanium is years behind POWER and SPARC64 right now, regardless of HP and Intel spin. Just look at the published benchmarks. The scalability is not even close when compared to Itanium... As far as SPARC64 vs POWER, on a pure speed basis, yes POWER is tough to beat (especially core-to-core), but when you're talking the high-end it is all about overall system. Throughput on the entire SPARC line completely blows by Itanium and POWER, especially when price is taken into account.
"Here we are, nine years later, and Sun's "Niagara" family of Sparc T servers are not competitive with the most recent x64 boxes."
Sun's processors have never been competive with Intel/AMD CPUs. However the systems have had consistently better IO, threading and reliability than most x86 systems. If the CMT chips are so uncompetitive, why the continued increase in volume sales of the T series boxes during the 25+% drop in other systems over the past few quarters?
And how come Sun got to 54%, by volume, of UNIX server shipments a couple of quarters back, according to Gartner? This could only be due to shipments of the cheaper T series systems as they were behind IBM and HP on revenues.
Author writes, "Here we are, nine years later, and Sun's 'Niagara' family of Sparc T servers are not competitive with the most recent x64 boxes."
The CoolThreads T family was built for web facing applications.
It took nearly 3 years for an x64 procesor to edge past a CoolThreads T1 processor's benchmarks posted in 2005. In mid-2009, there is still not a single x64 CPU which can outrun a single processor CoolThreads T2 benchmark posted back in 2007.
http://www.spec.org/cgi-bin/osgresults?conf=web2005&op=fetch&proj-COMPANY=256&proj-SYSTEM=256&proj-PEAK=256&proj-HTTPSW=256&proj-CORES=256&proj-CHIPS=256&critop-CHIPS=0&crit-CHIPS=1&proj-CORESCHP=256&proj-CPU=0&proj-CACHE1=0&proj-CACHE2=0&proj-CACHE3=0&proj-MEMORY=0&proj-NETNCTRL=0&proj-NETCTRL=0&proj-NNETS=0&proj-NETTYPE=0&proj-NETSPEED=0&proj-TIMEWAIT=0&proj-DSKCTRL=0&proj-DISK=0&proj-SCRIPTS=0&proj-WEBCACHE=0&proj-OS=0&proj-HWAVAIL=0&crit2-HWAVAIL=Jan&proj-OSAVAIL=0&crit2-OSAVAIL=Jan&proj-SWAVAIL=0&crit2-SWAVAIL=Jan&proj-LICENSE=0&proj-TESTER=0&proj-TESTDAT=0&crit2-TESTDAT=Jan&proj-PUBLISH=256&crit2-PUBLISH=Jan&proj-UPDATE=0&crit2-UPDATE=Jan&dups=0&duplist=COMPANY&duplist=SYSTEM&duplist=CORES&duplist=CHIPS&duplist=CORESCHP&duplist=CPU&duplist=CACHE1&duplist=CACHE2&duplist=CACHE3&duplist=NETTYPE&dupkey=PUBLISH&latest=Dec-9999&sort1=PEAK&sdir1=-1&sort2=SYSTEM&sdir2=1&sort3=CORESCHP&sdir3=-1&format=tab
For what the CoolThreads T processor are built for, they are very competitive, consistently being years ahead in throughput performance against X64 processor released.
You can always relie on Novatose! Even after the market has shown how little they believe the Sunshine he's out there plugging away, spewing out SPEC stats.
"....The CoolThreads T family was built for web facing applications......" Hold on a sec, is this the admission that Niagara is designed for the web niche that Novatose was so energetically denying earlier this year? Last time he was insisting it was perfect for enterprise apps like Oracle and SAP, but now he admits it's just a webserving wannabe. Problem with that admission is - although it's a very large niche - it means customers will look at the broad swathe of uses they can put their flexible x64 servers to, then look at the narrow spectrum of where Niagara makes sense, and they will largely go with the general purpose x64. If it was any otherwise you'd see web companies like Google throwing out all their cookie-tray x64 servers and buying nothing but Niagara, and we know that didn't happen.
".....For what the CoolThreads T processor are built for, they are very competitive, consistently being years ahead in throughput performance against X64 processor released." Which means they are completely irrellevant for 99% of the tasks customers buy x64 servers for, because x64 performs adequately to very well at a much lower price and with a much wider range of apps. The reality is Niagara is simply too expensive and too restricted to survive, especially in a downturn, even if it does outperfrom x64 in one niche area.
People like Novatose may still be daydreaming but I suspect Larry has already seen the light, and it ain't Sunshine. It's simple maths. Unless Larry can find a way to balance the books so that the money made by Niagara in its webserving niche exceeds the costs of developing, building and selling the servers then he will simply not do it. The only way for Larry to come even close would be a massive reduction in the Sun hardware business, which the cancellation of Rock was probably a precursor to, but he will have to cut a lot closer to the bone soon. Hence the rumours of the rest of the hardware bizz being offered again to companies like hp.
/SP&L
Whatever. Is "poor unemployed child" your nom de plume? Must be quite a few Sunshiners already out there looking for new jobs. But don't worry, all is not lost, the majority of Sunshiners can still be helped to learn an useable OS like Linux. However, I sense from the low quality of your response that you are one of those destined for asking "Would you like fries with that?"
Not much longer to wait now. Larry reckons he can tie up the whole deal by the end of the summer (not surprising really as the only sticking point seems to have been how many millions Oracle had to pay to get the Sun board to jump ship voluntarily), and if Soracle doesn't immediately roll out some convincing product roadmaps then the current customer defection rate will probably double. So we should all know pretty soon just how little of the bloated Sun carcass will get processed into something useful, how much will be sold on, and how much will just be discarded.
/put the popcorn on, it's almost time for the main event!
To use a term from TV, Sun jumped the shark with the StorageTek purchase. They have been dead men walking since then. 4$ billion wasted on StorageTek, poor management, inefficient internal systems (it takes them a wekk or more to get a quote to a customer), no leadership. Oracle cutting and slashing is likely too good for the middle and senior management of Sun. It's sad that they will walk away with millions.....
There has been very little posted on the openSolaris community site since December/January (did they lay-off those employees?).
I'm also surprised that their sales in second Q 2009 were not off by 50%, creative book-keeping?
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
