Sun Announces New x86 Servers
An anonymous reader writes "Sun announced the new V60x and V65x servers (1U and 2U respectively). The 1U has 2.8GHz Xeon CPUs and the 2U has 3.06GHz Xeon CPUs. They also announced a partnership with RedHat and Oracle running on these boxes. RedHat will also start shipping Sun's Java with their distribution."
I'm eagerly awaiting their move into adult markets with XXX86 servers. And Windows XXXP support!
Incidentally, Pr0stx0r fr1stx0r
Bitches
is this a ditch effort by sun to stay alive? it seems as though they've just been slipping away in recent years; they had the "One" platform of ubiquotous distributed computing and then that sort of disappeared, are we going to see the end of solaris soon?
From 68-248.sun.com
Remember when the x86 workstations began to eat into SGIs bottom line? They responded with building x86 workstations. The same thing is happening to Sun. Their SPARC servers are not keeping up with x86 servers, just as SGI IRIX/MIPS workstations began to lag in performance.
Now before the slashdot crowd begins to scream "But hey! The Sun Fire V480 is really fast!", remember that it is $19,995.00 in the base configuration. You'll get 10 IBM rack servers for the same price. In a clustered enterprise situation 20 3GHz Xeon will perform better than 2 900MHz UltraSPARC. Especially if we are talking Java.
Just as SGI was faster in the absolute high-end, so is Sun. The E15k is a monster. For some very specialized applications, this may be the only way to go. But for the very large majority of systems being purchased, a simple x86 server will do, especially if you can cluster it. This is where Sun is loosing the grip. Earlier you had to have a SPARC machine for advanced enterprise computing. These days are over, just as you had to have a SGI to run 3D software.
Now they are competing head to head with Dell in the x86 arena. This is a bold move. Wonder how long they will last.
Vintage computer adverts: http://www.vintageadbrowser.com/computers-and-software-ads
Wow.. I mean it worked so well for SGI, its a wonder everyone doesn't realease servers like this..
Having the latest versions JDK and also J2EE SDK built in with the system may mean that the Apache Tomcat/ant and other things will also come bundled with redhat (and most likely pre-configured just like mod_perl and mod_php).
A brand new installation of redhat can then run things like servlets, jsps, etc., just like we can now run mod_perl and all that without end users having to build and install it.
S
Let's see. Sun decides to release commodity hardware with the option to take commodity software, and charge a non-commodity price. So what makes Sun better than say, Dell, HP or the many other commodity vendors? Sun will be finished at this rate.
How is this going to hurt the MS strangle-hold over the desktop environment? Neither SUN or RH could compete with MS for the desktop; why would the two together be any different? If the two were to partner up and dump huge amounts of capital into developing a more robust and usable desktop there may be a chance.. But I doubt seriously that this is their intention. This may put a ding in MSs armor, but it won't affect their overall control of the desktop arena. But that's just my worthless $0.02.
"Reality is a crutch for people who can't handle drugs" - George Bernard Shaw (1856 - 1950)
"Free your OS"
IBM's 1U Server
Sun's 1U Server
At least they are price competitive with IBM. I'm not too sure about Dell but it's a start.
In Soviet Russia, Trojan exploits YOU!
Actually it was the incompetent management of Mr. Beluzzo who wanted to force feed NT down SGI's throats. SGI had long moved from the "graphics gravy train" and were making inroads in the Super and server markets. Mr. Belluzzo or whatever the spelling of his name is tried to move towards NT when SGI still had a technological lead, plust it made disasterours alliances with Microsoft (they gave tons of IP to Redmond for mere peanuts).
Incidentally Belluzo left SGI after almost running it to the ground, and jooined Microsoft right away, some people think that he was on M$ payroll even while he was destroying, er I mean managing SGI. Coincidende?
Every major company that has got in bed with M$ and based their business on NT offerings is either dead or dying: Intergraph, DEC, etc. etc..
RedHat will also start shipping Sun's Java with their distribution
w00t. One less thing I have to do after install.
Karma: The shiznight, mostly because I am the Drizzle.
Isn't this the same Redhat that refused to ship KDE because QT was not free software ?. When did java become Free Software ?
Wasn't it a well established fact that Except for SCO, none of the primary UNIX vendors ever developed a UNIX "flavor" to operate on an Intel-based processor chip set.
for the first time, apple to apple comparison shows sun cheaper than dell. i selected sun v65 and tried identical system at dell. dell doesn't give 3.06 GHz in 2U rack, so i selected 2.8 GHz. This is 600 cheaper. However, Dell charges $600 for upgrade from 2.6 to 2.8, so their upgrade from 2.8 to 3.06 would have been higher than 600 (upgrade from 2.4 to 2.6 is 200, 2.6 to 2.8 is 600). dell comes with customer installed RedHat Advanced Server while Sun comes with Solaris 9 and both are atleast comparable system (to be frank, RH profession is cheaper. but i am aware of several server apps which require RH AS patches and won't be certified on RH Pro).
Not in your lifetime. RH just isn't a good desktop distribution; Mandrake is much more polished and has fewer bugs. RH's real strength is in an enterprise envirionment. Similarly: Java is pretty weak for desktop apps (a survey of AWT, Swing and SWT should bear this out) but it's perfect for web interfaces and business logic.
The real fight is in the server world. Java + RedHat Linux is a winning combination, if they can get it right.
Finding God in a Dog
I just read the fine print... Suns JVM will only ship on Red Hat's Enterprise Linux Product.
Karma: The shiznight, mostly because I am the Drizzle.
I'm in a corporate environment where trying to get Linux or *BSD into the data center is an uphill battle. If the box comes from Sun, and runs Oracle, that makes that argument a whole lot easier for me. Even if it's more expensive than commodity hardware, they do have a deserved reputation for solid hardware, and I can use the logic that if Sun is willing to put their name on it, they're willing to back it up. I'm building a support system that's going to need it's own database; this box is worth looking into, for me.
"Hey, not enough people are buying our most profitable hardware! Lets give them MORE reasons to think about buying something else!!!" - Scott McNeely.
Captain Ahab is not going to do down with the ship until he has managed to feed the White(Wintel) Whale his entire reason for being. Thank you, Scott. We hardly knew ya. I hope you Sun employees out there know how to tread water, and while you are at it, try to keep those resumes dry long enough to get them distributed.
Of course, they will be trying to beat both Dell and IBM at their own game. SGI was at the last Linux Expo I went to (A few years back now) and during their presentation I was struck by the fact that they were trying to beat IBM at their own game, and I knew IBM was going to end up being the better player. Sun has more market share, extreme java expertise and a full range of machines to choose from, so I think they have a much better chance than SGI did.
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
Be mindful that Red Hat attempts to provide an array of software and services to their customer base - this includes the mainstream (read: free) distribution that contains GPL (or near-equivalent license) software.
Otherwise, Red Hat produces other distributions (like Advanced Server and Enterprise Server) that might contain proprietary (read: not so free) code and software that may require additional licenses.
The spirit is in open-source - but customer wishes also pay the bills.
Ayup
Sun-One as in Sun and their stock price a year from now...
Like it isnt easy enough to download it and unpack it on your own.....
While it is a very cool system... (incl. the 4 network interfaces), 1 GHz UltraSparc IIIis are slow, and they don't have the extra benefit of tons of cache compared to the regular US3. The RAM is only SDRAM (still), and 72GB of space is paltry.
So, if you absolutely need a SPARCv9 architecture rackmount, this is the way to go. But featurewise it falls short of say an Altus 140 from Penguin Computing, or even a 1000E if you want 64-bit. And Peng. Comp. is expensive as far as that kind of thing goes.
That being said, the small Enterprises are quite cool, but they aren't as cost effective. It helps if your organization has a pre-existing agreement, and can get you a break.
Black holes are where the Matrix raised SIGFPE
I agree 100% as much as I love to see mature open source databases with enterprise featuers I got to admit that it is still not coming anytime soon.
IIRC 6-7 months ago a marketing person from Oracle came to our company to discuss if Oracle will be suitable for our next development project, our customer contracted us to develop an online electronic components database with over +20 Million component with all their information, spec sheets. In the first year the database is expected to reach 0.7TB.
I recall asking her that we plan to implement heavy server side logic in stored procedures, and she said "if it couldn't be done with Oracle then it can never be done", you get to love Oracle's marketing people but technically she was right.
The only ready-today open source database that comes near Oracle (actually it is equivalent to Oracle 7) is SAPDB and what drives me really nuts is that the open source community ignores it completely in favor for something like MySQL (not that something is wrong with MySQL).
Erik
[1] http://sources.redhat.com/mission.html
Imagine a Beowolf Cluster of THESE!!!
Its interesting to think that it could be Sun's Java that is to blame for the demise of Sun hardware to some extent. Many server sales are now for people needing servers for J2EE applications which tend to scale out instead of up. Why spend significantly more on Sun hardware to scale out when you could just throw more X86 cluster nodes at the problem and achieve similiar results for less money.
OK, let's assume for a second you actually can build a rackmount server with exactly the same specs for less. Can you build 200 of them?
Sun isn't getting into this market to sell individual systems for people to run in their bedroom / offices. They are in the market to sell these in multi-rack installations to run the web servers that are attached to the clusters of Sun Fire 4800 app servers, which are attached to the failover-capable pair of 15Ks running Oracle on the back end.
Now, lets see you whip up a 72-way 576GB machine with over 100 GB/s of local memory bandwidth.
- Old Man of the Mountain ---- "I want to disturb my neighbor"
Hardware RAID. Dell sells Intel boxes with it, HP sells Intel boxes with it. Why should I consider Sun without this? I know Dell's Hardware RAID isn't the best performance, but it's great for availability--actually I prefer's HP's (er, Compaq's) RAID controllers the best. What does Sun bring to the table to compete?
I think this is actually a smart move on Sun's part. For several years, Dell, IBM, Compaq have been taking away market share at the low end from Sun. Now they at least have a product that can compete in the commodity space, where price is the overriding factor. It also means they have products all the way from the low end to the mainframe class like the E15k. Even if they only sell a small number of boxes, it will be better than they have been getting before.
As an IT manager, this is great, because I can buy Sun everything if I want to, which makes purchasing and maintenance easier. I can also push for a better volume discount if I want. Better still, at the low end, there's no vendor lock in, as I can run Open Source software under Linux. I also get the choice of Solaris x86 or Redhat. My experiences with the build quality of Sun equipment would give me some confidence too.
I think this is good for the customer and Sun. About time, I say...
are we going to see the end of solaris soon?
Absolutely not the end of Solaris. Sun is shifting some of their focus (if not most) from producing hardware to being a software and services company. Although this was announced along with the Red Hat deal, this is actually an attempt by Sun to compete Solaris against Linux at the low end. Sun is basically admitting and re-acting to what people have been saying for months (if not years) - Linux has been eating at Solaris by replacing high-cost sparcs with low cost x86.
The Red Hat deal is an obfuscation. The real aim here is to co-opt Linux by having current Solaris shops stay with Solaris. Lots of these shops that would have replaced the Sparc/Solaris platform with Linux are now going to be induced to stay with Solaris on x86. Sun figures that it is better to sell Solaris services without Sparc than to sell nothing at all.
Up until now, Solaris on x86 was always a "redheaded stepchild" at Sun. The hardware support was terrible and limited (very few video cards, for example). Hopefully, Sun will now give x86 good hardware support.
...in order to move a small production machine off aging hardware. The price is in-line with other 2U rackmount servers from Dell and IBM and probably a better deal than sparc hardware with the same specifications. The fact that we'd be running Intel + Linux on this machine as opposed to sparc + solaris is a huge benefit. That coupled with the Sun support services (which I think are prett good. Just like all support programs there are problems but as a whole its worth having) makes these new servers VERY attractive.
Now I have to convince my department to try this out...shouldn't be much trouble considering its $8,000.00
======== In the future, everything will be artificial. ========
And people don't need Oracle databases. Most of the free SQL flavors work just fine, thank you.
Give me PostgreSQL or MySQL any day over Oracle crap
Let me guess... you're 18 and haven't had your "introduction to Relation Databases" class yet, right? Kid, go away, you're bothering me.
Back in 1988, I remember seeing x86 based Suns running SunOS. It wasn't called Solaris back then. SPARC RISC based workstations weren't available then. The bulk of the Sun workstations were Motorola 68xxx based. Sun came out with an 80386 workstation called the 386i.
I had the opportunity to touch one when they first came out. A coworker was all excited that they were moving all their CAD software to the 386i, and he took me to their lab to show me the new machines. I wiggled the mouse, and it immediately crashed. That was the extent of my exposure.
As far as I can tell, the Sun 386i flopped. Linux was not around yet. SPARC came along a couple years later, and Sun migrated totally to SPARC. Perhaps their first attempt at x86 was a good idea, but poorly executed.
Hopefully, Red Hat will be shipping the JDK, not just the JRE. Generally, the only thing I have to do to a system after installation (aside from system updates) is installing Sun's JDK.
It's too bad Red Hat didn't do this previously as it would have saved people a lot of trouble (particularly when they didn't realize they were using kaffe rather than Sun for java).
-David
* As is generally the case, my opinions do not reflect those of my employer.
Sun already outclasses PC hardware with their SPARC systems. Xeons plugged into Sun's NUMA systems would be interesting but redundant.
These servers address those parts of a customer's application that may not need those spiffy and expensive busses. It is good for Sun to adequately address this part of the business. They need to keep their big box customers from getting distracted when shopping Dell or IBM for pizzabox application servers.
A Pirate and a Puritan look the same on a balance sheet.
Does anyone know if Java for solaris/x86 is better optimized than Java for linux/x86?
I know that Java for solaris/sparc has some specific garbage collection options.
SGI and DEC were going into the ground long before getting in bed with MS.
These companies were completely proprietary -- everything from the CPU to the Window Manager on their OS. Being proprietary is very, very, very expensive -- and the companies realized this. In the long term there was simply no way they would be competitive with Wintel/Lintel. So they did the only thing they could -- tried to leverage "commodity" technology onto their propreitary platforms and see if they could ride it out for a few years.
SGI also got royally screwed when Itanium was delayed for 4 years. They were left with shitty hardware while the highend got it's last hurrah boost during the dotcom years.
It's funny that IBM's and HP's "successful" Linux/Windows on Intel strategy is really no different than the failed strategies of SGI and DEC. The only reason is that IBM and HP had more money to play with until the market shifted.
If you look at Sun's press release about Red Hat you'll see that Red Hat will be including the JVM with their RH Enterprise Linux distributions ... not with Red Hat Linux, and that Sun will only be supporting RH Enterprise Linux. Why? Because Sun still won't license the JVM for redistribution. I'm not saying that Sun is wrong here (it's their toy, they get to choose the license), but this is what has been slowing the acceptance of Java on Linux with many developers. (Except for corporate Java developers -- they love it, and thus, so do I.)
Sun's trying to balance control of Java against market acceptance, and Solaris against Linux. Sun obviously thinks that anyone who wants Java for Linux will go to the effort of downloading it from Sun, while at the same time they get to differentiate Solaris from Linux by including Java. On the other hand, Sun could hardly sell & support Linux on Sun servers without also including Java; this agreement gives them what they want without letting go of their (perceived) control of Java.
We call it art because we have names for the things we understand.
A few years ago SGI decided that the Wintel market was where they had to be. Soon, they were peddling commodity Intel boxes running NT....and it was almost the death of the company.
I see Sun going down that road. Sun needs to learn from SGI's mistake. Their bread and butter was the high-end stuff. The stuff that makes scientists drool. These guys will pay anything for massive number-crunching ability. SGI realized this and decided to drop the low-profit commodity business.
Commodity Intel-Red Hat-Oracle boxes are NOT a way to build a profitable business.
-ted
Same reason the ultra-bulletproof zSeries are set up the same way. Shit happens, and when it hits the fan, you want a spare fan.
Does a Christian soccer team even need a goalkeeper?
Sun does not really compete at the low end. A 280R at 17,000 GBP (discount yadda yadda) is not really competitive with a Xeon whatever, but that's not the point. It is binary/OS/Firmware compatible with the F15K/6.8K that I will deploy on, and that's why I will specifiy if for a dev box every time.
Granted an x86 box will blow a 480/3800/280/240 into the weeds (probably) - but an x86 does not deliver the power of 72 CPUs on a 15K - or even 24 on a 6800. This is not about the back end - deploying a stateless web farm on x86 is cheap and good - but the back end DB/App server needs power (>24 CPU) and resilience (zero transaction loss fail over), and x86 does not offer the power at this stage.
Can you so SoLinux? How about SolarNux?????
... the new IBM PowerPC 970 may seem painful. But losing their high end business is probably EVEN MORE painful. Doing so would free up a lot of $$$ and perhaps make Sun more of an ally with IBM and perhaps even Apple.
Sun, more than ANY OTHER company is getting their ass kicked by Linux. It's simply no longer necessary to have a SunStation to drive a web server. Standardized PC hardware running Linux is MUCH, MUCH CHEAPER!!!!!
Sun's largest liability may be it's persistance in producing it's own CPU architecture. It may be VERY, VERY fast, but the marketplace for such a specialized platform is quickly evaporating.
Sun has created a sizable market for workstations in higher education environments. However, that could VERY EASILY ERODE with next generation purchasing. The push to standardize on hardware that can run EVERYTHING could very well doom Sun.
Sun needs to create an environment that their legacy scripts/software can run on PCs running Linux. Call it Solarnux if you will. That way they can sell Sun boxes that will fit easily into their traditional environments. At the same time these workstations could plausibly run Lindows to host Windows applications.
Similarly, their are lots of data centers out their that run on Sun hardware. Sheer cost decisions might force a platform switch. Linux keeps Sun in the game. The alternative is Windows2003 which puts Sun out of the ballgame. Why by a Sun box to run Windows2003?
At the higher end, Sun needs to seriously consider whether to keep making SPARC processors. A switch from SPARC to say
A switch of processing platform could bring a license to allow SunPPC workstations to run OSX apps. It could also bring an alliance to provide core services to creative shops that run OSX instead of Windows. In a way their fate is kinda linked.
Regarding SGI:
SGI better become a PURE SERVICES company or face certain extinction. LucasArts decision to buy Dell instead of SGI sounded the death bell for SGI. Their hardware business has been commoditized and they can no longer justify selling $25,000 workstations when a $3000 workstation with an nVidia Quadro card will do the same job.
-------- -------- Support Wesley Clark for president!!!
Unfortunately, the Sun Fire 2x0 are the only SPARCs with DDR.
The 440 and 880s use quad-interleaved SDRAM at (!) 75 MHz. I think they could ramp up that a wee bit... considering the RAM itself costs an arm and a leg.
Black holes are where the Matrix raised SIGFPE