X

What's 64-bit computing to Linux?

SuSE CEO Richard Seibt says the debut of 64-bit processors from Intel and AMD will break a logjam that has plagued IT server computing--with major implications for the ranks of Penguinistas.

3 min read
Unlike most areas of the technology business, 64-bit computing has somehow remained immune to the forces of commodity competition. Most 64-bit systems have historically been tied to proprietary operating systems. Compared with the more widespread x86, these systems only powered a small percentage of corporate computing shops and were relatively pricey. But with the introduction of its 64-bit Opteron processor line earlier this month, Advanced Micro Devices set powerful forces in motion.

First, some historical context is in order.

Itanium was the first operating system-agnostic chip with specific features to support Windows and Linux. The others were company-specific and thus, proprietary operating systems. Unfortunately for Intel, customer adoption went more slowly than the company would have liked following the Itanium processor debut in 2001.

That was partly because the market requires choices. As author and venture capitalist Geoffrey Moore noted on another occasion, markets can't fully develop in the absence of competition. The recent introduction of the AMD64 architecture with Opteron satisfied that requirement.

Without binding an operating system to their processors, Intel and AMD now present information technology consumers with a choice based on price-performance considerations. My guess is that this will free up the pent-up demand for 64-bit computing within enterprises. Big companies either are going to adopt or closely evaluate the new chips for future use. In a recent Evans Data survey of professional developers, 53 percent of the respondents said they considered 64-bit architectures to be important, while the same number of respondents indicated that they either are using or evaluating 64-bit platforms.

The availability of 64-bit Linux across all of these architectures makes for particularly interesting possibilities. In the enterprise, Linux is now routinely implemented for applications with higher importance on systems ranging from x86 PCs to IBM mainframes. In a CIO magazine survey of 375 information executives, 54 percent of the respondents said that open source would be their dominant server platform within the next five years.

These two trends--commodity 64-bit architectures and Linux--are intersecting. Five years from today, nobody in IT will be buying 32-bit servers (and maybe not even 32-bit laptops). They will buy 64-bit servers and almost universally run them with Linux. But in the technology industry, there is an old axiom that nobody buys computers--they buy application support systems.

So it is that a few other important software components must still become available before things kick into high gear.

All higher-importance IT applications are driven by databases, and thus software companies such as Oracle and IBM must deliver their database products for these newer 64-bit platforms.

My guess is that this will free up the pent-up demand for 64-bit computing within enterprises.
They, too, have been waiting for the 64-bit market to prove itself before making any major commitments. This should not take very long. Perhaps more than most other applications, databases benefit from 64-bit computing because of the need to directly address more than 4 gigabytes of memory.

Other independent software companies also must port their applications. These apps include popular enterprise resource planning (ERP), customer relationship management (CRM) and other application suites. Since all 64-bit server systems now have a single, unified Linux operating system, the porting procedure will be vastly simplified. This also will provide a boon to IT organizations that already use UNIX or Linux and develop in-house code. They will have fewer problems moving the rest of their computing infrastructure to 64-bits.

Might all of this help usher in the next generation of IT servers? That's something for which we technologists have been impatiently waiting. Now we're almost there.