News Stay informed about the latest enterprise technology news and product updates.

OSDL CEO: Microsoft has to accept Linux

Stuart Cohen of OSDL explains why his organization feels the "battle" between Microsoft and Linux doesn't have to be a battle at all. He also gives the skinny on some new features of the Linux 2.6 kernel.

Linux is an enterprise player, and customer demand will force Microsoft to make its products interoperate with Linux, says Stuart Cohen, CEO of Open Source Development Labs (OSDL), an organization that fosters the development and adoption of Linux.

OSDL recently rejected Microsoft's request for an independent study comparing Linux to Windows. Cohen believes Linux technology is much better than Windows, but he and others at OSDL believe that Linux-Windows interoperability, rather than war with Microsoft, is the way to go.

"Linux is legally sound, provides bottom-line business value and technical superiority," Cohen said, in an interview with He also talked about the new technologies coming in the new release of the Linux 2.6 kernel, virtualization and Linux, and the worldwide rapid rate of adoption of Linux 2.6.

The reality is that Microsoft customers are also Linux users.
Stuard Cohen,
CEOOpen Source Development Labs

Much is made of the "battle" between Microsoft and Linux. What's real and what's fiction?

Cohen: The reality is that Microsoft customers are also Linux users. The Global 2000 are running thousands of servers using both Windows and Linux. With Linux servers [placements] growing faster than Windows, the need for interoperability will become increasingly important. Because Microsoft listens to its customers, we expect it will look at effective ways to work with Linux and improve interoperability between the operating systems.

Can you shed some light on the controversy surrounding that Microsoft funded Linux vs. Windows study?

Stuart Cohen: I am still waiting to hear back from (Microsoft's) Martin Taylor on the results of his internal investigation at Microsoft about how the leak to the press of our private conversation came about. Until then, I'll stay with what I've already said. Why would OSDL want to participate in a study with Microsoft? Microsoft could probably find one negative line on Linux in a 100-page research report that it would spend $10 million marketing while ignoring the other 99 pages.

OSDL chief technology officer Tim Witham said in a recent interview that the new Linux 2.6 kernel will have improved I/O for RDMS, enhanced multi-path I/O for availability and more virtual machine capabilities. What else can you tell us about it?

Cohen: As Tim referenced, highlights include the improved I/O for RDMS and enhanced multi-path I/O for availability with virtual machine capabilities. We also expect updates to include 16-plus scalability, Linux security modules and multi-level security, Hotplug CPU and memory infrastructure, and support for IPV6, OpenIPMI, OpenAIS, OpenHPI and InfiniBand.

What's being done with scalable asynchronous I/O in the kernel?

Cohen: Scalable asynchronous I/O is being assessed. Before the feature is fully integrated into the kernel, there must be definitive performance improvements identified for a real-world workload.

How are decisions reached about what features to add to the kernel?

Cohen: Linus Torvalds and the appropriate subsystem maintainer decide what is added to the kernel. His decisions are influenced by the Linux development community and input from users, vendors, ISVs [independent software vendors] and a number of others.

Could you describe the struggle, if there is one, between the desire to add features to the kernel that enterprises need and to keep the kernel from being too complex, as some say Windows is?

Cohen: I would not describe this as a struggle. There is an ongoing discussion in the community with regards to benefit versus complexity. Regardless of the level of complexity, the community adds features when those features are proven to provide significant, demonstrated benefits.


Users: OSDL right to reject Windows/Linux TCO study

Novell idea man gauges the impact of open source

Virtualization is a hot-button topic with IT vendors, but are you seeing that much interest and/or adoption in the corporate marketplace?

Cohen: The focus on virtualization is an indicator of the maturity of the Linux marketplace and of Linux itself. That is, the need to balance and host existing loads already running on an OS and other applications for virtualization implies the commitment by those deploying Linux and the attitude of the marketplace backing Linux.

IT suppliers are responding directly to the needs expressed by corporate end users. OSDL hosts LUACs [Linux User Advisory Councils] around the world to collect input on requirements and best practices. LUAC members tell us that enterprise end users want to use virtualization as a means to build upon gains realized from their initial migration to Linux.

Most importantly, enterprise IT has migrated to Linux to leverage commodity hardware in lieu of legacy proprietary h/w running Unix. Linux lets these organizations first lower their hardware acquisition and maintenance costs for existing loads, and second, offers new options for horizontal and vertical scaling of applications and hardware supporting them. Virtualization carries this optimization one step further by fostering better h/w resource utilization through both load balancing (better virtual usage) and by supporting virtual clustering across CPUs (better horizontal utilization).

Has adoption of Linux 2.6 gone faster or slower than you expected? Are there geographical or market segment adoption patterns?

Cohen: Version 2.6 of the Linux kernel made its debut in December 2003. By the next summer, it had surpassed everyone's expectations by making its way into a range of enterprise distributions. By the end of 2004, most server and desktop platforms had embraced the new technology.

Currently, the only segment that has not migrated universally to 2.6 is the embedded segment, which is traditionally conservative. Adoption has been most aggressive in enterprise applications, e.g., communications infrastructure, while deeply embedded applications continue to build on 2.4 baselines. New distributions and new applications have been quick to adopt 2.6, especially to leverage features for real-time responsiveness and multi-processing.

Why has Linux adoption exceeded expectations?

Cohen: Linux is solving real business challenges for companies and organizations, large and small, around the world. It's important to assess, on an ongoing basis, what is making Linux a success so that others can learn from it and benefit from it. And, as Linux evolves to solve more business and technical problems, where does OSDL fit in its evolution?

The Linux development process follows the traditional scientific method with transparency, peer review and rapid innovation based on collaboration. As the world continues to 'flatten,' more and more countries, companies and individuals are looking for ways to meet their social, business and professional objectives. Linux enables low cost of entry for anyone who wants to participate in today's global marketplace. This is at the very crux of Linux success.

Dig Deeper on Linux servers

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.