This is a very interesting article that I got from the IEEE Spectrum, its about how Sun, Red Hat and Apple could combine and end the Microsoft domination...
Unite or Face Irrelevance
Why Apple, Sun, and Red Hat must merge
By Ralph Rodriguez
Here we are, almost three years into the downturn in technology stocks, and still nothing like a real recovery is in sight. To be sure, there have been a few interesting developments. Microsoft emerged from its antitrust trial last year with not much more than a slap on the wrist. And IBM embarked on its latest strategy to counter the Microsoft juggernaut, which involves embracing and marketing the Linux operating system as an alternative to Microsoft's ubiquitous Windows.
Here's another interesting turn of events: last May, Sun Microsystems Inc. (Santa Clara, Calif.) and Red Hat Inc. (Raleigh, N.C.), which markets a version of the Linux operating system, announced that they would distribute each other's products worldwide. It's a hopeful sign for the information technology industry, but not nearly grand enough. What we really need is for those two companies, along with Apple Computer Inc. (Cupertino, Calif.), to merge into a hardware-and-software colossus.
Granted, the idea sounds preposterous. But the technology sector clearly needs something spectacular to get back on the path to robust recovery, and it probably isn't another fight between Microsoft and IBM. We've been there and done that, over PC operating systems in the late 1980s and business software in the mid-1990s. And, lately, the skirmish between IBM and Microsoft has taken a nasty twist, as the two duel over intellectual property ownership issues involving a previously obscure company called SCO Group (for more, see http://www.osopinion.com/perl/story/21545.html).
It's hard to see how such machinations will do anything to help the technology sector. But think about the possibilities of a Sun-Apple-Red Hat merger. One of the brightest visions of the future of computing is built around the idea of grid computing, in which people will use the Internet to access geographically scattered computers, including extremely powerful clusters of computers. Compute power will become a utility of sorts, with users buying as much as they need for their tasks precisely when they need it. IBM, among others, is a champion of the approach. But a merged Sun-Apple-Red Hat would be in a position to offer a deeply compelling vision of grid computing, based on a sort of superplatform that could seamlessly span the grid, from the desktop all the way to blazingly fast Internet clusters.
The mammoth egos of some of the principals involved, to say nothing of the sheer complexity of the integration challenge, will probably ensure that no such merger will ever happen. That's a shame, because the alternatives are none too encouraging for Sun, Red Hat, or even Apple. Let's consider these companies' likely futures. Sun goes the way of the late, lamented Digital Equipment Corp., which reacted too late to the shift away from minicomputers and struggled in the 1990s and was ultimately bought by Compaq Computer. (Compaq was itself absorbed into Hewlett-Packard last year.) Not only will Sun go the way of Digital, but Sun's Solaris operating system could follow the trajectory of Digital's once-mighty Virtual Machine System (VMS). VMS lives on at Hewlett-Packard as OpenVMS, if you can call it living.
Linux goes mainstream
Red Hat has momentum, at least. The Linux operating system is the most popular example of "open" software; it is available free or at low cost and regularly refined by a community of like-minded expert volunteers. Based on the venerable Unix operating system, Linux logged some significant gains last year. IBM's embrace is to the tune of US $1 billion, to be spent mostly on support and integration of Linux with all its computer platforms. Big Blue's mainframes can now concurrently run multiple virtual Linux servers, greatly extending the mainframes' life and utility. Linux is the dominant operating system on Web servers.
Other signs of Linux's entrance into the mainstream include the introduction of a corporate-friendly version of the operating system by an alliance of four Linux companies (Red Hat not among them). But on the desktop, Linux still has few users in comparison even with Apple's operating systems, let alone Microsoft's. And Red Hat is beset by a growing horde of Linux providers, including the four that allied to produce the business-class version. Clearly, a merger with the larger, more established Sun and Apple would instantly elevate Red Hat above the minifray.
Apple made a splash recently with the surprisingly successful launch of an online music-retailing service. But becoming a niche player in the music-retailing business won't alter the fact that Apple is a niche player in the PC business. The company has never managed to convince more than a few percent of PC buyers to "think different." Despite its legendarily loyal following and solid products, it's hard to see how Apple will ever seriously challenge the Wintel juggernaut, which accounts for sales of almost all corporate PCs.
Linux at heart
What kind of a platform could Sun, Red Hat, and Apple create together? Let's say it's 2008 and you're a lucky EE working on an accelerator control module for a fuel-cell car. Your workstation is a neat little Apple Linux box capable of sustained processing rates in the neighborhood of 25 billion floating-point operations per second. Your monitor is panoramically wide, with enough room to display half a dozen applications, including real-time data screens, internally developed analytics, and standard Microsoft Office applications. It's the whole gamut of software available today, on one desktop.
Let's say you are running tests on your accelerator control module, which controls the flow of hydrogen and oxygen to the vehicle's fuel cell in response to movements of the accelerator pedal. You then decide to simulate the module, in real time with a stream of incoming data describing road conditions and driver actions. No problem: your computer is also part of an internally connected Linux cluster of processors acting as a single multiprocessor system. At the moment you need it, and transparently to you, your desktop system taps into the power of a score of other machines in your department.
--------------------------------------------------------------------------------
It's 2008 and your workstation is a neat little Apple Linux box capable of sustained processing rates in the neighborhood of 25 billion floating-point operations per second
--------------------------------------------------------------------------------
Now suppose you want to take your simulation up to the next level. You want to see how the accelerator controller affects the levels of power from the fuel cells, the transmission system, and the torque delivered to the wheels and the car's traction control system. These other electronic and mechanical systems are being designed by your co-workers in the cubicles around you. You tweak your circuit and see, more or less instantly, how it affects the other subsystems. So from the division management point of view, all the desks are for once truly interconnected. These computers can share not just data, as in today's networks, but also processing resources. They can also combine their resources to create a true division-wide computing entity.
This kind of sharing is already occurring in Linux clusters at universities and research centers all over the world. And this ever-expanding network of clusters is also being connected via the Internet into large grids of computers. That the future belongs to Linux clusters is suggested by the fact that in the most recent survey of the world's most powerful computers (see http://www.top500.org), a Linux cluster placed fifth—ahead of many conventional supercomputers that cost 10 times as much. The cluster, which has 1152 processing nodes, was built at Lawrence Livermore National Laboratory in California and achieved a processing rate of 5.69 trillion floating-point operations a second in a benchmark test.
It's easy to see what Red Hat would contribute to this hypothetical merged computer company: it is the Linux expert. But what would Sun and Apple bring to the party? Sun's role would also be straightforward. It has more skilled Unix developers and engineers than any other company on the planet, and the most advanced and reliable multiprocessor servers that run on Unix. The clincher is that Sun also has unparalleled expertise in the Java language, often used in Web-related applications.
Unite or Face Irrelevance
Why Apple, Sun, and Red Hat must merge
By Ralph Rodriguez
Here we are, almost three years into the downturn in technology stocks, and still nothing like a real recovery is in sight. To be sure, there have been a few interesting developments. Microsoft emerged from its antitrust trial last year with not much more than a slap on the wrist. And IBM embarked on its latest strategy to counter the Microsoft juggernaut, which involves embracing and marketing the Linux operating system as an alternative to Microsoft's ubiquitous Windows.
Here's another interesting turn of events: last May, Sun Microsystems Inc. (Santa Clara, Calif.) and Red Hat Inc. (Raleigh, N.C.), which markets a version of the Linux operating system, announced that they would distribute each other's products worldwide. It's a hopeful sign for the information technology industry, but not nearly grand enough. What we really need is for those two companies, along with Apple Computer Inc. (Cupertino, Calif.), to merge into a hardware-and-software colossus.
Granted, the idea sounds preposterous. But the technology sector clearly needs something spectacular to get back on the path to robust recovery, and it probably isn't another fight between Microsoft and IBM. We've been there and done that, over PC operating systems in the late 1980s and business software in the mid-1990s. And, lately, the skirmish between IBM and Microsoft has taken a nasty twist, as the two duel over intellectual property ownership issues involving a previously obscure company called SCO Group (for more, see http://www.osopinion.com/perl/story/21545.html).
It's hard to see how such machinations will do anything to help the technology sector. But think about the possibilities of a Sun-Apple-Red Hat merger. One of the brightest visions of the future of computing is built around the idea of grid computing, in which people will use the Internet to access geographically scattered computers, including extremely powerful clusters of computers. Compute power will become a utility of sorts, with users buying as much as they need for their tasks precisely when they need it. IBM, among others, is a champion of the approach. But a merged Sun-Apple-Red Hat would be in a position to offer a deeply compelling vision of grid computing, based on a sort of superplatform that could seamlessly span the grid, from the desktop all the way to blazingly fast Internet clusters.
The mammoth egos of some of the principals involved, to say nothing of the sheer complexity of the integration challenge, will probably ensure that no such merger will ever happen. That's a shame, because the alternatives are none too encouraging for Sun, Red Hat, or even Apple. Let's consider these companies' likely futures. Sun goes the way of the late, lamented Digital Equipment Corp., which reacted too late to the shift away from minicomputers and struggled in the 1990s and was ultimately bought by Compaq Computer. (Compaq was itself absorbed into Hewlett-Packard last year.) Not only will Sun go the way of Digital, but Sun's Solaris operating system could follow the trajectory of Digital's once-mighty Virtual Machine System (VMS). VMS lives on at Hewlett-Packard as OpenVMS, if you can call it living.
Linux goes mainstream
Red Hat has momentum, at least. The Linux operating system is the most popular example of "open" software; it is available free or at low cost and regularly refined by a community of like-minded expert volunteers. Based on the venerable Unix operating system, Linux logged some significant gains last year. IBM's embrace is to the tune of US $1 billion, to be spent mostly on support and integration of Linux with all its computer platforms. Big Blue's mainframes can now concurrently run multiple virtual Linux servers, greatly extending the mainframes' life and utility. Linux is the dominant operating system on Web servers.
Other signs of Linux's entrance into the mainstream include the introduction of a corporate-friendly version of the operating system by an alliance of four Linux companies (Red Hat not among them). But on the desktop, Linux still has few users in comparison even with Apple's operating systems, let alone Microsoft's. And Red Hat is beset by a growing horde of Linux providers, including the four that allied to produce the business-class version. Clearly, a merger with the larger, more established Sun and Apple would instantly elevate Red Hat above the minifray.
Apple made a splash recently with the surprisingly successful launch of an online music-retailing service. But becoming a niche player in the music-retailing business won't alter the fact that Apple is a niche player in the PC business. The company has never managed to convince more than a few percent of PC buyers to "think different." Despite its legendarily loyal following and solid products, it's hard to see how Apple will ever seriously challenge the Wintel juggernaut, which accounts for sales of almost all corporate PCs.
Linux at heart
What kind of a platform could Sun, Red Hat, and Apple create together? Let's say it's 2008 and you're a lucky EE working on an accelerator control module for a fuel-cell car. Your workstation is a neat little Apple Linux box capable of sustained processing rates in the neighborhood of 25 billion floating-point operations per second. Your monitor is panoramically wide, with enough room to display half a dozen applications, including real-time data screens, internally developed analytics, and standard Microsoft Office applications. It's the whole gamut of software available today, on one desktop.
Let's say you are running tests on your accelerator control module, which controls the flow of hydrogen and oxygen to the vehicle's fuel cell in response to movements of the accelerator pedal. You then decide to simulate the module, in real time with a stream of incoming data describing road conditions and driver actions. No problem: your computer is also part of an internally connected Linux cluster of processors acting as a single multiprocessor system. At the moment you need it, and transparently to you, your desktop system taps into the power of a score of other machines in your department.
--------------------------------------------------------------------------------
It's 2008 and your workstation is a neat little Apple Linux box capable of sustained processing rates in the neighborhood of 25 billion floating-point operations per second
--------------------------------------------------------------------------------
Now suppose you want to take your simulation up to the next level. You want to see how the accelerator controller affects the levels of power from the fuel cells, the transmission system, and the torque delivered to the wheels and the car's traction control system. These other electronic and mechanical systems are being designed by your co-workers in the cubicles around you. You tweak your circuit and see, more or less instantly, how it affects the other subsystems. So from the division management point of view, all the desks are for once truly interconnected. These computers can share not just data, as in today's networks, but also processing resources. They can also combine their resources to create a true division-wide computing entity.
This kind of sharing is already occurring in Linux clusters at universities and research centers all over the world. And this ever-expanding network of clusters is also being connected via the Internet into large grids of computers. That the future belongs to Linux clusters is suggested by the fact that in the most recent survey of the world's most powerful computers (see http://www.top500.org), a Linux cluster placed fifth—ahead of many conventional supercomputers that cost 10 times as much. The cluster, which has 1152 processing nodes, was built at Lawrence Livermore National Laboratory in California and achieved a processing rate of 5.69 trillion floating-point operations a second in a benchmark test.
It's easy to see what Red Hat would contribute to this hypothetical merged computer company: it is the Linux expert. But what would Sun and Apple bring to the party? Sun's role would also be straightforward. It has more skilled Unix developers and engineers than any other company on the planet, and the most advanced and reliable multiprocessor servers that run on Unix. The clincher is that Sun also has unparalleled expertise in the Java language, often used in Web-related applications.
Comment