"It's Too Late! CIOs Don't Care Anymore"

The future of Unix Part 1: IBRS' view: Here's Kevin McIsaac of IBRS responding to a question about Sun's port of Solaris to x86/x64: "But too late! Linux has come along, Windows has come along and it runs really well on Intel. CIOs don't care anymore. There is a lot of great stuff about Solaris but guess what? The mainframes had a lot of that for twenty five years and I left the main frame to go to Solaris back in the eighties. I think organisations like Sun are kidding themselves. They should know from their historical roots that good enough trumps best and Linux and Windows are for most people, good enough!"

Again with the "it's too late" bit. I haven't seen this one crop up in a long time, though. And this time it comes with an exclamation point for extra special emphasis! My goodness. When people say it's too late for what we are doing (Solaris 10, OpenSolaris, new x86/x64 systems, new SPARC systems, etc) that only demonstrates that they are ripe for being disrupted. They are not paying attention. It's not too late. The timing is perfect.

And anyway, we don't 'port' Solaris to x86, we develop simultaneously on both platforms. Sparc and x86 are on equal footings. And the rest of the article isn't much better either, it's mostly uninformed drivel.

Posted by Alan Burlison on October 02, 2007 at 12:39 PM JST #

Thanks for the correction, Alan. And I agree. The rest of the article is quite bad as well.

Posted by Jim Grisnazio on October 02, 2007 at 12:51 PM JST #

I had this really nasty discussion on our local Linux Users Group list the other day with regards to Linux. Someone who is rather prominent in the Linux community declared that "Linux is READY for the desktop" to which a few of us said "uh...no it isn't."

One of the reasons it is not is because "mainstream" or generally accepted software providers that work in the corporate space don't support Linux on the desktop. Why? Mostly because there is almost no Linux on the desktop in corporate environments. Software manufacturers won't waste the development dollars on something that will not return their investment.

When I was with CA, our CEO John Swainson and a few others met with Scott McNealey who got beat us up on not supporting Solaris 10 on both Sparc and Intel. While CA committed to do so in the end, the reason there was no support yet is because there were VERY few CA customers running Solaris 10 on Intel.

It all comes down to market share. One of the reasons Linux became so popular is that Sparc systems are quite expensive as compared to Intel systems. Linux was / is free, and the community rallied behind it. Red Hat came on strong as did Novell. The message about replacing Sparc servers quickly spread, and thus did the support for Linux.

My point in all of this is that unless the Sun marketing machine and product management jumps on the enterprise evangelization train NOW, it might be too late.

Posted by David Meyer on October 02, 2007 at 02:39 PM JST #

The "Good Enough" argument works when "Good Enough" is significantly less expensive than "Great". Mincomputers were significantly less expensive than Mainframes. RISC/UNIX systems were were significantly less expensive than Mincomputers. Linux/x86 systems were significantly less expensive than RISC/UNIX systems.

Obviously, Sun's decision to offer a version of Solaris for x86 with a pricing model competitive with the Linux pricing model is disruptive. No sane person should buy "Good Enough" when they can get "Great" at the same price.

The real reason many customers continue to choose Linux rather than Solaris for migrations from SPARC to x86 is simply fashion. It is exactly the same reason certain styles of clothing are popular today. Everyone wears them because they are popular, and they are popular because everyone wears them. IT has been a fashion industry for years, because most people working in IT so fear failure they choose paths which provide an excuse if the fail. "Nobody ever got fired for buying IBM." "We chose product X because it was in Gartner's Magic Quadrant." "Everyone is using Linux." Rationality often does not come into play. Realize Sun benefited significantly from this in the late 1990s, when customers literally thought their IT projects might fail if they did not use Sun.

The biggest problem I see with Linux in the Enterprise is the management of kernel patch levels with kernel level drivers. For example, look at the twenty-eight different drivers for Oracle's Automatic Storage Management required for various SLES 9 x64 kernel patch levels: http://www.oracle.com/technology/software/tech/linux/asmlib/sles9.html.

For the most part, Linux distributions get around this by incorporating these drivers in their kernel. But for those which are not, think about the effects of just patching the kernel in a Linux environment, or the dependencies matrix for a system build.

This is area where Solaris has distinct advantage. Most Solaris device drivers work not only through patch levels, but through major releases. For example see this Emulex driver: http://www.emulex.com/emc/support/sol_sparc.jsp.

Posted by Mark on October 02, 2007 at 04:39 PM JST #

Oh, crap Jim, I just realized this is that Windows fanboy Kevin McIsaac from Meta back in the 1990s. I remember a Meta customer who almost shifted their planned PeopleSoft implementation at the last minute from Solaris to Windows because of that.

"Well [IBM] offer[s] both. At the end of the day, they're a very different company. IBM doesn't rely on its proprietary hardware revenue."

Really Kevin. Are you saying IBM doesn't rely on revenue from its Mainframe and POWER systems line? I believe the Mainframe is IBM's most profitable product, and in CY2006, the POWER series drove close to $4 billion in revenue.

"They have got AIX, the T-Series, the mainframe and they've got all their storage kit. They've got all this Intel gear, they used to have PCs and now they've got services."

I assume "T-Series" is a typo. Anyway, I do know IBM's xSeries x86 server group is not highly profitable like software and Mainframes. Nobody is making a lot of money selling x86. Nobody. And IBM's storage is not taking the world by storm, and neither is HP's or Sun's.

"What's Sun got? Sun's got Solaris and it's got Sparc. They don't have anything else that makes money really."

Now Kevin is using the perceived profitability of a company's division to make claims against a company's strategy.

In other words, Kevin believes companies should offer x86 with Windows and Linux, as that is the future. This is exactly what he said in I think, 1999, but Kevin was clueless about Linux, and he only spoke of Windows. But if you do exactly what Kevin says, and offer x86 with Windows and Linux, he discounts it because it is not big enough. He will not commit to a definition of success, and he will move the goalposts. All because he desperately wants to stay right.

His entire self-esteem is still wrapped up in that paper he did for Meta almost 9 years ago. How pathetic.

I will agree IBM has a very good business model, and very good portfolio management. IBM does not rely solely on its proprietary hardware revenue. It also relies on its proprietary software revenue (DB2, WebSphere, & Rational). It's all in IBM's annual reports. But honestly, I don't think Kevin knows much about how IBM actually works.

Oh, and where is "Part 2" of this series, with the Sun interview? Apparently it does not exist, at least it does not show up on a search of the web site or on Google.

Posted by Mark on October 02, 2007 at 05:10 PM JST #

Post a Comment:
Comments are closed for this entry.


« April 2014

No bookmarks in folder