IBM Doubles Rewards For Ditching Sun

Taking advantage of the uncertainty surrounding Oracle's acquisition of Sun, IBM has doubled the monetary incentives they are offering to ditch Sun gear. Offering $8,000 in software or services for every Sun Sparc processor ditched for an IBM Power server, the program seems to be paying off. IBM has helped 1,640 customers migrate from other manufacturers' hardware over the last year. "The program applies to Sparc-based Sun hardware, such as the Sparc, UltraSparc, and Sparc 64 servers, and also to Fujitsu systems that run on Sparc chips. A customer that moves off a Sparc-powered system running, say, eight processors would be eligible for up to $64,000 worth of rewards."

Comments

The Death of SPARC?

[eldavojohn]

Is this the death of SPARC?

I would have said murder but I'm not interested in a hardware flame war. I mean, I know Fujitsu and some lesser known companies are using it but I'm not sure in what capacity. Is this the end of SPARC?

Can any hardware experts comment on whether or not this is the end of this architecture? Or does it have some niche market/capability like PowerPC?

I guess OS support could have been a cue that it was on the way out but is there any reason to be concerned that it's apparently done?

Re:The Death of SPARC?

[teflaime]

SPARC was a dying hardware platform anyway. Sun was shipping far more Intel product than SPARC. It's too bad. SPARC was pretty good for the level it was designed to operate (mid-range area). IBM and HP have somehow convinced everyone that P5/6 and Itanium somehow fit in that environment, but they are really out of the price range and overpowered for those needs.

I'm just hoping Solaris survives the Oracle take over. I still like Solaris better than Linux for webservers and such, personally.

Re:The Death of SPARC?

[fm6]

Sun was shipping far more Intel product than SPARC

I work at Sun on x64 products (both Intel and AMD) and this just isn't true. The x64 products are doing well, but our sales are still predominantly SPARC. The long-term strategy has always been for Sun to place more emphasis on x64 products, but not to the exclusion of SPARC systems. And so far, x64 hasn't even achieved parity with SPARC, or anything like it. Why? Not something I'm going to comment on in a public forum.

A lot of Slashdotters seem to think that Sun has turned into a kind of white box server vendor. Even if we we totally abandoned SPARC, that's not going to happen. Our market niche is high-end computing, and always has been. In the x64 world, it means that in order to compete we have to do stuff that white boxes can't. This includes fancy lights-out remote management, really high computer density (anybody else have an 4U system with 8 processors and a half-terabyte of RAM?) and a greener machine with few plastic parts and a lot of power-conserving measures. These things require a lot of clever engineering, and are the only reason we have any successful x64 systems at all.

I have no idea what Oracle has in mind for Solaris. Contacts with them are, if anything, more circumscribed than they would be under normal circumstances. But in my own inexpert opinion, it's not a coincidence that we've been acquired by one of the few software vendors that's still serious about Solaris/SPARC application.

Re:The Death of SPARC?

[ILongForDarkness]

At my work we have a couple dual socket T2 sparc servers (T5140) that we are using as fileservers for 30 disk arrays, 150TB of disk space. We went with them because we liked SAMFS (Sun's hierachial storage management system), and the T2 chips have 8 cores X 8 way threaded for a total of 128 simultaneous compute threads in a 1U server.

They can push a lot of I/O(60GB/s of I/O bandwidth per chip) but I wouldn't want it for compute intensive stuff because they only have 1 FPU per core, and 2 integer units per core (ie 8 threads have 1 FPU and 4 threads have 1 ALU). Anyways the current generation seemed to be targeted at I/O intensive stuff especially highly threaded protocols(eg. samba/NFS).

Re:The Death of SPARC?

[turgid]

I used to work for Sun too, just about the time they acquired Niagara from Afara (IIRC). Tremblay left Sun to found a startup which designed multi-core, multi-threaded CPUs, and they came up with Niagara, which was basically 8 very simple UltraSPARC cores on a chip, each core capable of holding 4 thread contexts which could be switched in and out to hide memory latency.

This is how things used to work at Sun. Every so often, very clever people with "lunatic fringe" ideas would leave to found startups with VC money to realise the ideas, and Sun would buy them back when it looked like it might work.

Sun's in-house CPU design is pathetic, which is why UltraSPARC started to lose out to x86 in the late '90s. Given the size of AMD's team and their complete lack of funds compared to Sun, Sun should have had much better CPUs than Opteron/AMD64, but look what happened. Fujitsu did much better with SPARC64.

Buying MySQl was a bone-headed decision which finally killed Sun. They tried to buy a name for over $1bn and got nothing. As always happens with these take-overs, the lead developers left. Remember the Cobalt purchase? What about StorageTek? Are any of them left?

There were many opportunities Sun should have taken but didn't. For example, they should have bought AMD right when Opteron came out (but Not Invented Here! - and it took some pretty loud shouting to get the Solaris prima-donnas to get Solaris on Opteron) and given AMD the task of developing UltraSPARC along side Opteron. Heck, some of us wondered, if a 64-bit x86 can be made to go so fast, what would it be like if the x86-translation layer was replaced with a SPARC-V9 translation layer? BIG HINT.

Now, calling GNOME the "Java Desktop System" was suicide. Potential customers were saying, "Why would I want a desktop written in Java?" Marketing PHBs, I hope you have learned a lesson!

Why did they ditch a bunch of the standard applets and rewrite less featured and slower ones in Java? I remember seeing a 1-pager proposing to write an MP3 player in Java for the JDS. Meanwhile we were shipping xmms on the Companion CD!

Java, Java, Java, Java, Java,..... Blah!!!!

PHBs, you may be interested in Java, but that was not Sun's core business, despite what you wanted to think, and all it did was alienate the millions of dedicated Unix people who used Solaris and Sun gear.

So I bought some Sun shares in the employee discount scheme. When the takeover stuff was announce, the shares almost went up enough that if I sold them next year when I don't have to pay any tax on them, I'll only have lost about 25% of my money.

I could have run that company better... Every year, buy some other companies with a few thousand employees, have a RIF and get a write-down against tax. Great strategy guys.

Phew, I needed that. End of rant.

Ditching Sun servers

[OolimPhon]

Does this mean that there will be a market full of cheap(ish) second-hand Sun servers your average geek might be able to make use of?

Did IBM really want to buy Sun . . . ?

[PolygamousRanchKid]

I never really believed it. But the "due diligence" gave the opportunity for IBM to take a peek at what Sun has underneath its fingernails.

Sun is down on the ropes, and IBM would like to give it a knock out.

Yeah, IBM might have wanted to control Java, but the hardware . . . they've got enough hardware of their own.

I'm not switching...

[Anne Thwacks]

We use excludively OpenBSD on UltraSParc servers for our financial transactions processing. I am not switching - I want uptimes of a year and I certainly dont want to port our software to another OS or hardware. $8k wouldnt go near that. (We have over 20 CPUs, but porting is not going to happen while my Sun kit works). I have never paid Sun a penny for support. Their kit is reliable.

What a load of old Cock & Bull

[RotateLeftByte]

I work on MQSeries and have been involved with message queueing systems since 1982.

WMQ is very reliable and has been since V5.1 came out. Pretty well every large financial organisation in the world uses it to move trillions of $$$, ££££, Yen, Euros around their companies & beteween them on a daily basis without error.

Please backup your statement with a list of 'Showstopping' bugs in WMQ.
And no (before you ask), I don't work for IBM.

Re:Most of them...

[gfxguy]

IBM isn't going to re-sell the Sun hardware.

Huh...

I was tending to agree with the notion that it wasn't anti-competitive until I ran across this line.

If they are not going to resell it, then they are taking a trade in value that they cannot recoup. It's not quite, but it seems similar to dumping (lowering prices below your costs in order to drive a competitor out of the market).

I'm not saying I have a problem with it, but look at it this way... if I was disinclined to by an IBM before, and IBM offered to sell to me below cost (by giving a way overpriced value for a trade-in) in order to get me to switch, I think that'd be considered anti-competitive.

Re:Most of them...

[ThePhilips]

The CoolThreads stuff is neat, but never really took off at the volume Sun was hoping.

It's way too Java biased.

My company did evaluation (C/C++ stack of applications) with only one result: disabling the CPU multithreading capabilities improves performance. Otherwise, sustained performance penalty makes the whole solution not worth its money.

But it would be interesting to see how they do Java applications. In our stack, Java has bits of business logic, but mostly (one of the) front-end(s) for customers to hook up their own applications - it's not performance critical thus was not evaluated.

As for POWER...there will always be people who need the biggest, fastest, baddest processor. A lot less people need them than they used to - x86 commodity keeps getting faster. But there will always be the top X% of the market that needs speed. That's why IBM sells POWER. And hey, while we're catering to them, we can also use it in our run-of-the-mill servers (AIX, AS/400, Mainframe, etc.)

I think POWER has a lot more staying power than SPARC.

I always had the opinion that IBM keeps POWER floating simply because it's pretty much always delivers profits. The POWER among architectures is like Linux among OSs: it scales from embedded systems to clusters to mainframes. IBM is pragmatical company and POWER apparently sells well: there seems to be undying demand for custom chips for all possible applications. Provided simplicity of POWER (some folks do implement PPC32 of FPGA) IBM can very quickly adjust it to requirements of a customer. If one market is slowing, other markets do support future development.

In contrast, SPARC to be profitable has to have a wider market: it's not that flexible accommodating various application fields. They are present in essentially one (huge) market: servers. Yes, it's huge - but highly competitive market. In past, the sole reason for people to buy SPARC was Solaris: good, stable - slow - yet best server OS. I'm not sure how that would play out after Oracle's take over. Seeing how now Oracle runs on Linux/x64, I have strong feeling that SPARC might be the first business unit sacrificed. Solaris/x64 (which Sun already produces) might be way too tempting for Oracle as a way forward.