Sun Unveils 64-bit Server Line 287
SumDog is one of many to let us know, PC World is reporting that Sun is expected to reveal the first few of their new 64-bit servers at their quarterly product rollout. From the article: "Formerly code-named Galaxy, the Sun Fire X2100, X4100, and X4200 servers represent the company's bid to woo customers, particularly the financial industry sector, away from rival server vendors Hewlett-Packard and Dell."
And it's based on Opterons... (Score:3, Informative)
Should we take this as the final sign that Sun is giving up on Sparc?
And as they move toward "normal" chips, should we expect that Sun will be able to continue to offer the hardware advantages (say, to do with reliability) that they held with Sparc, or are we going to be seeing them move closer to being a plain-box Opteron reseller-- in the same way that as Apple is moving to plain-jane x86, they are also giving up on technologies such as Open Firmware?
Sun is not giving up on SPARC (Score:5, Informative)
These new servers absolutely rock, and at superb prices.
I once had the pleasure of a 4-way Opteron v40z with a development version of 64-bit Solaris 10. It was a screamer, especially compared to our 4-way Dell P4 Xeon box, and 64-bit.
It was plenty fast enough to host 4 zones and several developers working on KDE, gcc and all manner of other stuff.
At last, Sun looks like it's turning the corner (despite the best efforts of some of its PHBs - no names mentioned).
Good luck Sun.
Sun should transition away from SPARC (Score:2)
Re:Sun should transition away from SPARC (Score:2)
Niagara is an interesting engineering research project and a first attempt at multithreaded CPUs.
What do you know about ROCK that we don't? Englighten us.
Re:Sun should transition away from SPARC (Score:2)
All I know is what I read in the papers, and what Sun has been saying is '30x USIIIi @ 1.2HZ' shipping in 2008 at the earliest. By 2008, that kind of performance isn't going to be very impressive as we will probably see 4 core, 65nm X86-64 chips from both AMD and Intel by then.
Re:Sun is not giving up on SPARC (Score:2)
How often would you like your bank to switch the OS on the systems that manage your account transactions?
Re:Sun is not giving up on SPARC (Score:2)
Re:Sun is not giving up on SPARC (Score:2)
(a) You're smoking dope, and (b) have not a fscking clue as to the complexity of converting a large commercial system from one platform to another.
And no, migrating a DNS or Web server from Windows to Linux isn't a 100th the scope of migrating a bank from AIX to Solaris.
Re:Sun is not giving up on SPARC (Score:2)
That's what my roommate does. He works for a company that writes banking software, doing the conversions for banks moving from other systems, frequently running Solaris or AIX, sometimes running Linux, to theirs, which runs on Windows, which is a hell of a lot more difficult than switching between Unices.
Maybe you thought I didn't read the GP post, but I did. I didn't mention a damn thing about DNS or web servers; I was talking about banking software. Pull your head out and realize that k
Re:Sun is not giving up on SPARC (Score:2)
Re:Sun is not giving up on SPARC (Score:2)
Re: (Score:2)
Re:Sun is not giving up on SPARC (Score:2)
Linux the kernel supports SPARC, but so what? It's apps that are important, and not all apps run well on Sparc. Has to do with gcc, I think. The debian-sparc ml would give more detail.
Re:Sun is not giving up on SPARC (Score:2)
Re:Sun is not giving up on SPARC (Score:2)
Re:And it's based on Opterons... (Score:5, Informative)
Re:And it's based on Opterons... (Score:2)
Re:And it's based on Opterons... (Score:3, Insightful)
Re:And it's based on Opterons... (Score:2)
HP DL 760 [hp.com]
Re:And it's based on Opterons... (Score:2)
HP DL 760
i'm not sure why you are posting links to servers with list prices of approx $40k.
Re:And it's based on Opterons... (Score:2)
The HP DL 760 is not intended to compete with Sun's $750 (diskless) x2100.
Re:And it's based on Opterons... (Score:2)
Re:And it's based on Opterons... (Score:2)
Hardly! See Rich Teer's blog [blogspot.com] for more information on Niagra or The Register [theregister.co.uk]. Apparently the GHz ware is finally is over.
yes.... (Score:3, Funny)
Sun 10 years from now (Score:2, Interesting)
Re:Sun 10 years from now (Score:5, Informative)
It depends on whether you think Sun is turning the corner with these new servers. The original opteron line was basically a company on life support getting pretty much reference models out the door. While these machines show Sun's polish all over it. I think these servers compete well with HP and Dell's offering and they have Sun's polish. I am hopeful. But ten years is a long time from now.
> Will they survive by selling 'mostly' software?
Huh? This is a server line that runs Solaris or Linux. They are definitely still selling hardware and giving away the operating system.
> I know they sell hardware, but they no longer control the full stack like IBM with POWER.
Sun has almost never had control over the full stack. They sold you the hardware with a free (as in beer) operating system on it. Then you put on the application/server software. They might help you buy that application/server software. But they have never made it.
Re:Sun 10 years from now (Score:2)
Huh? There's the Java Enterprise System software stack. Many of the components, including the app server has been around for a while.
Re:Sun 10 years from now (Score:2)
Funny, I used to work for Netscape. Sun & AOL carved us up between them, and Sun kept the server stacks. Netscape begat iPlanet begat SunONE begat the current "Java Enterprise System".
Their J2EE appserver has been completely re-written, but their web & directory servers are derivatives of the old Netscape codebase.
Re:Sun 10 years from now (Score:3, Informative)
Huh? This story is about a new line of servers and youq uestion if sone is selling mostly software!?!?! And you get modded interesting. I think it's pretty interesting that someone thinks it's a valid question.
These boxes are completely designed by Sun. Though the CPU is not manufactu
Re:Sun 10 years from now (Score:5, Informative)
Re:Sun 10 years from now (Score:2)
Re:Sun 10 years from now (Score:3, Interesting)
Actually, I do see them surviving (I don't know about the 10 years, it's a long time), but not selling software. Neither do they, otherwise they wouldn't have taken the decission to open source their OS and provide it as a free download. They seem to even be planning to give away their hardware.
I think they are realizing where the real money is and moving to a subscription model. You get powerful
Suns have been 64 bit for a while now... (Score:5, Insightful)
Re:Suns have been 64 bit for a while now... (Score:2, Insightful)
Re:Suns have been 64 bit for a while now... (Score:5, Insightful)
That's a great reverse-reverse-psychology troll or something, but it's hardly insightful. I have always been a Sun supporter over the decade or so I've been working with Solaris on Sparc. I have always said that they were the best *nix thing going out there, compared to their competition. But I was also always a fan of Linux where it was applicable.
But the time came (some time ago now) to admit that Sun has in fact missed the boat on Linux. What made it especially frustrating was that, of all of the commercial *nix vendors, Sun was in the best position to capitalize on the Linux wave. They were already all about developing and promoting open standards (think NFS and NIS back in the day, among many others). They were already the best non-free platform to build and use open source software on. Hell starting with Solaris 8 they were shipping a good deal of open source software with stock Solaris. But some idiot(s) in charge of the company completely lacked the vision to make it happen. I can only imagine how much better a position Sun would be in (and how much better off all consumers of *nix would be) if Sun had re-centered themselves around Linux kernels going forward back in the late 90's or even 2001-ish. They could've turned their kernel engineering teams to work on Linux on Ultrasparc (and Opteron), and could've brought a lot of scalability and other enhancements with them to the Linux kernel in general to boot.
Even now that Sun has started to turn the corner on Linux from their previous stances (which were to ignore it, and then to marginalize it as a toy), their stance still smells a lot like, "Sure, run linux on our Sun-branded but otherwise whitebox-like and overpriced x86 and x86_64 hardware, but only for crappy unimportant edge devices. Leave all the real computing to a real operating system like Solaris." Meanwhile smart companies are working out strategies to transition off of the last remaining Ultrasparc behemoths they have left in the corner of the datacenter while the majority of their real computing is already happening on Linux today. Average not-so-smart companies will be doing this in a few years.
I don't hate Sun, and I don't think they're Evil. But I think someone fell asleep at the wheel there and completely failed to take advantage of the Linux wave like Sun should have. If anything, I feel sad for them, it's tragic to watch a great company go down like this. They could still turn it around, but I don't have much faith anymore that they will.
The Linux Boat? (Score:3, Insightful)
What "wave" would that be? Sun already had a Unix at least as good as Linux. Face it, the only thing Linux has going for it is support for all sorts of hardware. Other than that, Linux can be supremely annoying -- no manpages, for one thing. Add the Solaris 10 features like
Re:Suns have been 64 bit for a while now... (Score:5, Insightful)
Why would the company with "the best *nix thing going out there" abandon that just to jump on the linux bandwagon?
It sounds like you're saying you wish they had done that so that they could contribute "scalability and other enhancements" to the linux kernel for the public, but what would that gain them? Hell, what if that alleged better scalability is based on the fundemental design of the kernel? Suns engineers may not be able to just write some "scalability modules" and plop them into Linux.
And if in fact their own kernel was already better than linux's kernel in that regard, again why would they want to abandon their own kernel? Just to say "Hey, look everyone, LINUX! LINUX! SQUAWK!"
None of this is a slam against linux, which is part of a perfectly fine OS as well. I just don't buy into the argument that linux is the ultimate end result of all OS evolution. Just because linux is good doesn't mean it's what everyone else has to be. Sometimes, different is even better than good.
Come on, give people more credit then that.. (Score:3, Funny)
If not, maybe you shouldn't be reading Slashdot. It's too technical for you. Go read C|Net.
Sky banners (Score:3, Interesting)
Haven't seen one of those in ages
Spells Death for the SPARC (Score:4, Interesting)
The marketing talking head will claim that SPARC lives in Niagara and Rock, but note that Intel is now building a new x86-64 implementation that focuses on multicores just like Niagara and Rock. Given a choice between Niagara/Rock and Intel's/AMD's new multicore chips, most customers will prefer the latter.
The only future remaining for the SPARC is in esoteric highend systems built by Fujitsu and destined for simulating weather, nuclear explosions, and overpopulation.
Re:Spells Death for the SPARC (Score:2)
Re:Spells Death for the SPARC (Score:2, Interesting)
Sun has started down a long road of realization that their proprietary chips aren't worth the silicon they're printed on in a marketplace that values interoperability over nearly all else. HP realized this a while back, and have all but phased out their Alpha and PA-RISC lines.
That said, there is a place for non-x86 chips. HP has replaced most of their Alpha and PA-RISC lines with x86 chips, but some of the high-end boxes went to Itanium. IBM is still pushing POWER -- hard. I don't sus
Re:Spells Death for the SPARC (Score:2)
Re:Spells Death for the SPARC (Score:4, Interesting)
When the Niagara based servers are released we'll have a 1 socket, 8 core, 4 threads per core server which in a 32 server rack gives us a 1024-way grid in one rack
Now, the Niagara CPUs performance (specifically floating point performance) is lower than Opteron (Sun have made no secret of that) but for heavily threaded/moderate computation workloads, a grid of Niagara CPUs looks like a very interesting proposition.
Rock is rumoured to be SMP capable so rather than building grids of these boxes running seperate OS instances you are able to build 1024-way (maybe more) SMP servers with significantly less power consumption and much higher performance/watt and performance/$ than existing SMP (from ANY company) in a footprint that is a fraction of the size of current highend servers and mainframes.
I wouldn't say SPARC is dead yet, it might have been a bit ill for a while there, but it's on it's way back.
Re:Spells Death for the SPARC (Score:3, Insightful)
Re:Spells Death for the SPARC (Score:2)
Re:Spells Death for the SPARC (Score:3, Interesting)
-Jem
Re:Spells Death for the SPARC (Score:2)
Linux based Itanium servers from SGI are the biggest Unix servers on the planet.
128 cpus? Why think so small?
Re:Spells Death for the SPARC (Score:2)
You forgot about Cray. IIRC, they sell systems with hundreds of Opterons...
Well yeah, Cray sells you XT3 systems with up to 30000 Opterons if you fork over enough cash. But those are distributed memory computers. The parent poster fawning over big Sun SPARC:s probably meant shared memory computers. In this class the winner is SGI with currently a maximum of 1024 (or was it 2048) CPU:s running a single Linux kernel.
Re:Spells Death for the SPARC (Score:2)
high-end servers will always need serious superscalar RISC processors.
What does "superscalar RISC" have to do with SMP scalability?
Re:Spells Death for the SPARC (Score:3, Informative)
When I typed the original message I forgot about Itanium2, so being RISC is not a prerequisite for massively parallel systems. All of the above are superscalar designs.
-Jem
Re:Spells Death for the SPARC (Score:2)
You keep using that word. I do not think it means what you think it means.
http://en.wikipedia.org/wiki/Superscalar [wikipedia.org]
Re:Spells Death for the SPARC (Score:2)
Re:Spells Death for the SPARC (Score:3, Informative)
The Opteron supports glueless 8-chip systems. Just wire the HyperTransport links together and off you go. Of course, with dual-core that's already 16 CPUs, and will be 32 CPUs next year. And it's quite possible to add bridge chips to support more than 8 Opterons.
All of the above are superscalar designs.
Including the Opteron.
Entire Cray XT3 line (Score:2)
http://www.psc.edu/publicinfo/news/2004/2004-10-2
Re:Spells Death for the SPARC (Score:2)
If I were Sun, I'd be asking AMD very nicely to make an Opteron processor with a SPARC V9 translation layer on it instead of an x86 one.
In fact, if Opteron were anything like the Transmeta processors, AMD should be able to put multiple translation layers on the chip. It would be totally and utterly cool if it could run AMD64 and SPARC V9 in hardware, on the same processor and be software switchable. Yo
Review over at Anandtech (Score:5, Interesting)
Re:Review over at Anandtech (Score:2)
Miracle machine! (Score:3, Insightful)
The two hard drives can be setup for RAID 0, 1 or 10 via the BIOS.
Now, it may be a few years since I took classes in college that touched on various RAID levels, but one thing that I DO remember is that RAID 10 requires a minimum of 4 physical drives...
X4100 Review at InfoWorld (Score:4, Informative)
Favorite line... (Score:4, Funny)
Re:Favorite line... (Score:3, Funny)
What I want to know is, what the heck were they doing paying attention to a bunch of grandmothers? Most of them can't even use a VCR, much less Solaris...
No, no, no, no. (Score:5, Insightful)
Where bold insert Customer
That's simialr to Digital's downfall. They built some of the best computers in the world, thinking if we build it they will come. But it wasn't what the customers wanted. The same goes for catering to Wall Street. They want short term quick earnings growth; not necessarily long term custoemr growth. That may not be be conducive to achieving a product line that will last and the customers will even want.
Re:No, no, no, no. (Score:3, Informative)
v20z/v40z? (Score:2)
The v[24]0z servers were not manufactured by Sun themselves, and they've mentioned that they're working on their own version.
I love my v20z servers - they are a great alternative to the crap Dell calls 1U servers. I hope these are as good, and maybe a little better supported.
How about 64 bit java? (Score:5, Interesting)
Re:How about 64 bit java? (Score:2, Informative)
is this what you meant?
Let's hope they run better than the W2100z WS (Score:5, Informative)
4 BIOS updates later and the problems have dwindled a bit but we constantly get BSP error messages on boot up and random DIMM error messages during POST (on both sockets and chips that have been thoroughly tested and known to be good). Daughter processor cards have been bad as well (already replaced 4 in a batch of 40 which, according to Sun is "acceptable rate of failure").
Their latest BIOS update (version R01_B4_S2, released last month) does resolve the frequency of some of these errors but now we have machines that lock up on that BIOS release but not previous ones.
I only post this because the chips are Opteron 250s by AMD (64-bit) and the main board is another AMD.
Based on my experience with these workstations I wouldn't touch anything put out by Sun until they can get a quality control department set up and running anything with AMD chips.
Re:Let's hope they run better than the W2100z WS (Score:5, Funny)
You obviously don't know how to admin Linux you fucking cluebie, go back to Windows. Oh, Solaris? Sun sucks, use Linux.
+3 Funny
Re:Let's hope they run better than the W2100z WS (Score:2)
But these problems are mostly during POST so the OS has nothing to do with it.
By the way, Red Hat Enterprise Linux 3 is our OS of choice on these machines. We're staying away from Solaris.
Re:Let's hope they run better than the W2100z WS (Score:2, Informative)
They should... they are entirely different boxes. The new ones are from the acquisition of Kealia (Andy Bechtolsheim's startup).
Re:Let's hope they run better than the W2100z WS (Score:3, Informative)
Microsoft Windows is fully supported by Sun, too (Score:5, Informative)
Local Windows admin was impressed (Score:4, Interesting)
For example, one of the Windows admins here got a 1U loaner Sun box running Windows {something} Server. (I don't remember which specific version.) He was very impressed by the speed and stability(!!!) of the system. Being a Sun admin for over 10 years, I, of course, had to bust his chops about the Sun logo on the box and "upgrading to a better operating system." That's when he told me that it ran Windows.
They have a great marketing opportunity: a highly-optimized system that can run not one, not two, but three operating systems! Not only that, it will run all three of them well! Sun also gives a three-year warranty on their hardware. Most of the other systems that I've seen require you to pay extra for a 3-year contract.
Although I know that many will look at this as "moving to the Dark Side", I don't see a problem with this personally. It gets Sun in front of people that otherwise would not have looked at their hardware. Maybe - just maybe - that will help to broaden Sun's customer base, which can only help in the long run if Sun plays their marketing cards correctly. After all, their current business model is to sell the hardware, but they'll be glad to throw in the OS for free. So, they're not looking to make money off the Windows install. They're looking to make money because they got a sale that otherwise would have gone to HP/Dell/other.
Who knows? In the future as hardware progresses such admins might say, "Well, we have this Sun box that doesn't really do anything now. Let's download Solaris and see what it's like." Of course, I'd rather have them say, "Hey, you want this? We don't use it anymore..."
woo customers? (Score:3, Funny)
So these severs will be faster then most intel based processors with a lower price tag?
Didn't think so.
"Galaxy" class systems (Score:4, Funny)
Our sunfire was one of our worst investments. (Score:2, Insightful)
We had a sunfire as a mail server for about 3 years then went to upgrade the disk space last year and a 75gb drive was $4000. Proprietary isn't worth it.
Re:Our sunfire was one of our worst investments. (Score:2, Insightful)
I really hate this "proprietary" phrase getting thrown around with regard to Sun. Forget that they're using AMD CPUs now (the
Re: (Score:2)
and they'll be running Windows... (Score:2, Interesting)
Healthy competition (Score:2, Interesting)
UltraSPARC was not the first 64-bit CPU. (Score:3, Informative)
They were definitely preceeded by the DEC Alpha.
I am confused: why post these old news? (Score:2)
Niagara is much more server-oriented, while Opterons are more adequate for workstations and some types of server workload. But Niagara is much mroe suited for typical web serving and database hosting.
Re:I am confused: why post these old news? (Score:2)
Re:I am confused: why post these old news? (Score:2)
1 GHz is plenty if your target application can be multithreaded.
But in the end, we'll see what the market wants. Certainly there is a market for Opteron, Niagara and SPARC64 servers alike.
Galaxy? Does Sun have no organizational memory? (Score:3, Insightful)
"How do you make your Sun server run at 1/4 speed?"
"Add 3 more processors"
Airplane circling Dell with a message from Sun (Score:2, Informative)
Bah! 64-bit is for kids! (Score:2)
Re:specs available? (Score:3, Interesting)
Care to show any statistics to back that up?
Re:I love the bit in the article (Score:4, Informative)
Re:I love the bit in the article (Score:2)
hey no offense, but... (Score:5, Informative)
Suns don't run Windows and they don't run Linux
Actually, these new machines run Solaris, Linux and Windows - they are even on WHQL. They are the second-gen of Sun's AMD based x86-64 machines, and there were some intel x86-32 based systems before that, so arguably they are on their 3d or 4th gen of machines which can run Windows, if you like.
Compare and contrast this with Sun and HP who basically say "service, hey, you bought it, the check cleared and if it stops working then come see us about a service contract (which we will charge you up the wazoo for)".
Sun always quotes multiple service contract prices right there on the web page when you order the hardware (different levels of service).
Re:I love the bit in the article (Score:2)
Re:I love the bit in the article (Score:2, Informative)
are you on crack?
The Galaxy boxes run Solaris, Linux, or Windoze.
The current Opterons do as well.
RTFA.
why is gross misinformation being modded up as Interesting???
Re:I love the bit in the article (Score:2)
Re:I love the bit in the article (Score:2)
Who Modded This Up? (Score:2)
Re:USB 1.1 only? (Score:2)
That's fine for kbd and mouse, but nothing else. On a server what are they good for?"
You seem to have answered your own question, exactly. Of what need is there for USB 2.0 on a server?
Had the network interface only been 10/100 you would have a valid point, but on a server, USB isn't terribly important.
Re:USB 1.1 only? (Score:2)
Re:64-bitness (Score:2)
Except for the LX50 which was Pentium III.
so why even mention the 64-bitness? ctually, it would be more a news if sun were to release a 32-bit server.
It's news to the Windows-toting PeeCee people who are still buying 32-bit Pentiums from intel, which is 80% of the audience here, and 99% of Western business.