hi all/Peter,

"Migrate Solaris to Linux" : Solaris is still quite popular, although it might NOT stay that way by the end of next year if strategies don't change along with the landscape. How come you're NOT asking HP-UX/AIX to also do the same OPENLY ?

Atleast, solaris is OpenSource(HP-UX/AIX aren't). What sun needs to do is to add GPL to Solaris' license. That will ensure that it can quickly grab all of Linux's userland tools AND hardware compatibility to make it much more usable in the x86/EMT64 bit land.

SPARC : What happens to SPARC is a different question, though -- I don't have many thoughts on that. Althought it does NOT make much sense for SUN to help migrate to Itanium since it does NOT offer Itanium boxes.

LAMP : neither IBM/ORACLE have such a big stake in java as sun did (and does especially now). Java's strong point is the web side -- the only weakness being sun's SERIOUS lack of working relationships with small/medium ISPs to host java apps. If MS can work with 1000s of ISPs to make ASP hosting so easy, I don't see a reson why SUN can't do the same.

ALSO,

a) sun needs to add more consultants and find more 3rd party consultants who recommend/suggest SUN hardware/hosting options.

b) Solaris needs to play well with all the Virtualization environments (MS/VMWARE/XEN) -- so sun needs to collaborate with these guys.

c) MARKETING : very few people know what products sun offers -- how many know about the IDENTITY server, integration server, etc. They aren't doing enough!!

d) They're not using the java support channel to their advantage. only recently, they made it easy for customers to get support for java.

my 2 cents.

BR,
~A

Poor Jonathan! It's a full-time job just to read the advice that the blogosphere is offering, these days.
Peter, do you have any support for your claims that say "and it has not cannibalized their xyzzy sales"? I don't have IBM's internal financials, but I'd lay a good bet that AIX is selling more poorly post-Linux than it was pre-Linux.

Mike -

AIX has actually been selling very robustly, per the IDC UNIX report last year:

IBM claimed the number one position in global UNIX-based server revenue share, gaining 6.9 points YTY, up 31.8 percent YTY.

Last year both IBM and HP beat out Sun as the #1 UNIX vendor for the first time, years _after_ they both bet big on Linux/x86 (IBM with xSeries and HP with Proliant). Here are the 2005 IDC UNIX market share stats: IBM=31%, HP=30%, Sun=29% Go figure.

Mike,

You don't need IBM's internal financials - try IDC. Servers running AIX are now the #1 in Unix market share by revenue - when Linux was made available on Power, IBM would have been #3.

The question isn't about share -- it's about revenue. If the total Unix market has shrunk, and AIX has shrunk somewhat less, its share will have increased while its revenues declined.
Whether or not one product has cannibalized revenue for another is very hard to say conclusively, since you can't run a controlled experiment. I'd still bet large that some of IBM's customers who would have chosen AIX went with Linux instead, so it's a good thing that IBM has embraced Linux. Clearly Peter and Andrew have data before them, and I'm only arguing from gut.
Sun's transforming itself into a new kind of business, and every one of these changes is fraught with risk. It might not work, and it might reduce the value that the market places on the company. There is no question Jonathan has been bold. I'd argue that Scott was bold lately, too. The question is whether the boldness will be sufficient.
I'd rather be us, offering the advice, than him, living with the outcomes! I'm a big fan of Sun for a lot of reasons, and I hope that those outcomes are great.

Java already does everything PHP does and ten times more. Why would you need PHP if you have Java? I don't see the need to incorporate a weak non-object-oriented scripting language into an excellent platform. I don't agree with the first two points, but I completely missed the third point...

If you think SPARC and Solaris are dead, why not read "http://www.scalewithrails.com/downloads/ScaleWithRails-April2006.pdf"

Operating System choice is more varied than it has ever been. How can you say that is is a commodity. It now is a choice, and Linux is never always the answer. It is well down my list if I want to run a business on it.

This whole article smells of an IBM beatup.
P.S. I have used there Identity server for a very long time now. It works well. Never needed the integration server.

Yes, Sun needs to do better in the services area.
Knowing people who have worked for IBM GS, supporting Solaris has been very profitable for IBM.

Solaris v. Linux: things like DTrace exist because Sun owns their OS source and can put all the hooks into place. Won't work for an open-source OS. Linux doesn't need indemnification, there's no SysV code in there, sorry SCO.

Hardware: Sun is a *HARDWARE* company; the OS is just there to pull you in. Same thing as Apple. The new multi-core chips kick serious tail. Don't write the hardware off just because it can't run Windows.

LAMP: They'd be shooting themselves on the L and A part of that: SunONE/iPlanet/Netscape webserver isn't huge in the market, but the places you find it are usually Huge Businesses. And Java isn't a big money-maker, but they'd be chewed out for "abandoning" it if they started pushing PHP.

Peter/All, Sun's biggest mistake in the past 10 years has been its failure to capitalize (make money) on a successful technology it invented - JAVA.

So even though open-sourcing Java is going to help the growth of Java itself, how is it going to help SUN out, who had already invested billions on it. Someone has to correct me on that.

Solaris is NOT the problem, SPARC is. Companies replace Soloris machine not because Linux is superior, but because Linux/x86 is much cheaper than Solaris/SPARC, and deliver similar performance. The world does not need SPARC, or another cool design chip. Economic benefits of mass production far outweigh any technical disadvantage.

Java does not need PHP to help fix its problem. Java just need to evolve faster to catch up with C#/.NET. Programmers know what I am talking about.

My 2 cents for Jonathan:

Hire a top level Executive from Microsoft or Oracle or IBM who knows how to make money in software industry, and a tech Executive who can oversee development of software that will make money. Build up its consulting service to sell a whole package of integrated solution with Opteron + Solaris/Linux + Java Apps + Storage

What Sun needs to do is get better sales and fulfillment people. These are the frontline group that can truly hear what clients want. Sun's use of VARs does not allow them to penetrate SMB markets as Microsoft does. If JAVA had better RAD tools which would compete with Visual Studio at an "ease of use" level, instead of relying on XXX number of different tools, like eclipse, JAD, Net-beans, ANT, crap-more-crap-even-more-crap, then programming and acceptance would be even greater. Microsoft makes this extremely easy which is why developers choose it, hence, platform sales. Its a bit of a ground-up approach.
Migrate to Linux is a stupid approach. At a large scale, when you need absolute stability and TOOLS to monitor system stabilty, AIX and Solaris beats others hands down. Tools like Tivoli, Openview, Veritas, BMC Control-M and CA products enable stability and monitoring. Linux is a step in IBM's stategy, as an introductory OS. Then as you grow, move into AIX then at the largest scale of transactions, ZOS. Do you truly trust a LAMP stack to run your equity trading backend systems? Or order entry and reservation system? The major banks use LAMP on the peripheral but not at the core, airlines too.
Another possibility that Sun should investigate is to gradually divest itself from the SPARC fabs. AMD will tell you how much it costs to own and operate a fab, as will IBM. The current Fujitsu-Sun partnership for the SPARC chips is nice, but Sun should gradually get out of owning the fab business, even though the SPARC chip is a superior 64 bit processor.
The arguments on the page are for whether SUN is a hardware or software company. At the hardware level, not many are going to beat Dell. And as Chinese companies start offering cheaper systems and gradually buy US companies (think Lenovo), what would become of SUN if they were a hardware company? They should compete as a software and services company and possibly look into buying either a popular middleware company, like BEA, or investing into a database company or project like MySQL or Postgres to complete their "stacks".