In previous LWM articles, we have discussed the Open Source Development Labs (OSDL), Carrier Grade Linux (CGL), and the momentum Linux is gaining in telecom. We have also covered the current technological trend of companies moving away from proprietary technologies towards open and standardized platform components (hardware, operating system, and middleware).

Telecom platforms have very stringent requirements related to availability, performance, security, and reliability. To help telecom platforms move away from proprietary operating systems, OSDL has created a working group, Carrier Grade Linux, to specify requirements and subsequently help create open source projects to implement the missing features in Linux, thus meeting the CGL requirements. The result is the availability of a choice between a proprietary operating system and a Linux kernel with carrier grade features.

We interviewed Stuart F. Cohen, the CEO of OSDL, to get the full story on CGL, including the involvement of Linus Torvalds, the activities around kernel testing, the Linux Legal fund, and more...

LWM: Can you please introduce us to the Open Source Development Labs? Its history, members, mission, and working groups along with their goals.
Stuart F. Cohen: OSDL was founded in 2000 by IBM, HP, CA, Intel, and NEC to solve shared challenges in the Linux industry. Today, we have more than 40 members all over the world and employ several core developers such as Linus Torvalds and Andrew Morton. Our mission is to be the recognized center of gravity for Linux - the central body dedicated to accelerating the adoption of Linux for enterprise computing.

OSDL provides two testing facilities for Linux developers - one in Beaverton and one in Japan. In addition, OSDL working groups are a central place for vendors, customers, and developers to come together to improve Linux. OSDL currently has four working groups, dedicated to accelerating the adoption of Linux in the communications industry, in the data center, and on the desktop. In addition, OSDL has a localized working group in Japan pooling resources to address technical issues and features unique to Linux in the Japan market.

LWM: Carrier Grade Linux has been the most visible working group at OSDL. Can you please tell us about what CGL is trying to achieve and how?
Cohen: In a nutshell, CGL will help Linux become the OS used in the communications industry, while still yielding the benefits of open source.

CGL is made up of carriers, network equipment providers, software distributors, and software developers and is chartered to gather requirements, establish roadmaps, publish specifications, define architectures, and sponsor the open source projects needed to achieve a true Carrier Grade Linux.

LWM: CGL has recently released draft sections of the 3.0 specification for public comment. What are the goals of the 3.0 specification, and what can you say about the roadmap beyond 3.0?
Cohen: 3.0 is focused on enhancing the reliability and the cluster-ability of carrier-class platforms that are built with CGL. In addition, the CGL working group has taken another look at security for 3.0 and is adding a comprehensive set of features to address the evolving security needs of carriers using open systems. It's too early to get specific about future specifications since 2.0 is currently being adopted and 3.0 will likely be released later this year. The working group is looking at interfaces with middleware as well as application portability as goals they would like to pursue in the future.

LWM: Security features in communication environments are core requirements. To what extent is CGL addressing security issues?
Cohen: Carrier Grade Linux Requirements Definition version 2.0 (CGL 2.0), announced in October of 2003, continued the great work of improving general capabilities of Linux in the communications industry, but also focused heavily on two new areas: security and clustering.

CGL 2.0 addresses a number of security issues, ranging from better password protection to better tracking of log files and the ability to track suspicious activities.

At the end of the day, CGL will only be successful if it represents what communications companies demand - and security is a big issue to them.

LWM: The SA Forum and OSDL are both defining cluster usage models. Are these efforts synchronized to avoid redundant work or inconsistencies?
Cohen: Absolutely. OSDL's interest is to accelerate the adoption of Linux. Any time we duplicate another organization's efforts, it means we aren't using our resources appropriately.

LWM: Are there any commercial products that are available in the market and are validated to be CGL compliant?
Cohen: Distributions including MontaVista Software, Turbolinux, SUSE, and Miracle Linux have all announced support for the CGL 2.0 requirements.

Almost as important, there has been a marked increase in the number of communication RFPs that require CGL compliance. This is certainly a strong validation of the working group's efforts. Among the leading network equipment providers who have announced public support for the latest CGL requirements are Alcatel, Cisco, Ericsson, NEC, and Nokia.

LWM: Are there any test suites, which demonstrate that commercial products are CGL compliant? Or when will they be available?
Cohen: The CGL working group has developed a mechanism to allow distributions to register that their products are CGL compliant.

To satisfy CGL 2.0 priority-one requirements, Linux distributions will self-register against the OSDL Registration Requirements defined by the working group and post their results publicly on the Web. Posted results must include a declaration of how each priority-one definition is met. OSDL will also provide a list of those companies who complete the registration process with links to their Web sites.

We are actively working with network equipment providers (including Alcatel, Nokia, Cisco, etc.) who are using CGL-based distributions and the development community to make registration valuable and easy.

LWM: What benefits does OSDL foresee for adopting CGL as the operating system for communication platforms?
Cohen: Certainly, moving to CGL from a proprietary OS can save network equipment providers (NEPs) money because they don't have to develop, maintain, or license an in-house proprietary OS. Instead, they can invest in the CGL ecosystem to make Linux good enough for their own use. The CGL 2.0 requirements are prioritized based on input from leading network equipment providers, including Alcatel, Cisco, Ericsson, Fujitsu, NEC, and Nokia. Based on their input, Linux is becoming the OS of choice for secure, scalable, and reliable communications needs.

In the communication industry space, or ecosystem, OSDL's CGL working group serves as a catalyst for improving both the kernel and the software stack used in Linux distributions.

LWM: What role do open source implementations of CGL play in the wider adoption of Linux?
Cohen: While as a specification CGL has been targeted to meet communications industry's requirements, the specification is very viable outside that segment. CGL enables Linux to achieve very high levels of availability and reliability and these capabilities are needed throughout enterprise environments. And as it is a free specification, we openly encourage anybody to adopt it.

LWM: When will we see a large number of platforms running CGL?
Cohen: It's hard to say in a nutshell - although we do know that there is a lot of momentum. There are now several distributions which are CGL v1.1 compliant and several NEPs have announced CGL-based platforms.

LWM: We see that there are some common interests between Data Center Linux and Carrier Grade Linux working groups, in areas such as clustering, storage, and security. What is the interrelation between CGL and DCL? Is this targeted somehow within OSDL?
Cohen: DCL has a much broader target audience than CGL and a much broader directive. Issues such as security, storage, and clustering are highly relevant to all of OSDL's working groups. Although the CGL and DCL are separate working groups, we are working to share information so that resources are used most effectively.

LWM: How can people get involved with OSDL working groups, as companies and as individuals?
Cohen: Our working groups are open to any company or individual developer who wants to contribute. Companies who join gain voting rights that help steer the direction of the working group. Individuals with a bona fide interest in supporting the technical activities of the subgroup can join our working groups as well.

LWM: The use of Linux is being expanded day after day; Linux is running on desktops, servers, mobile phones, personal digital assistants, and on other specialized devices. As a general observation, if a feature is not widely used, or if it does not benefit almost everyone using Linux, then it is unlikely to be integrated in the kernel. The concern is whether there will be a fork of the kernel source tree into specialized kernel trees targeting different markets. If this was the case, who will be responsible of new kernel tree(s)? How will this affect the open development process of the resulting specialized kernels?
Cohen: Part of OSDL's mission is to accelerate the adoption of Linux. As a part of that, we feel that our role is to help prevent the forking that hurt the adoption of Unix.

LWM: OSDL is now home to Linus Torvalds and Andrew Morton. Was this a PR move or are they (Linus and Andrew) expected to contribute to some of the work being conducted within the OSDL working groups?
Cohen: OSDL is a natural place for Linus and Andrew to work. For Linus, this was an opportunity for Linux development to become his day job - in a vendor-neutral environment. Linus, Andrew, and other kernel developers at OSDL are aware of the working group activities and output.

LWM: What are the ongoing activities in OSDL in regard to kernel testing?
Cohen: OSDL is extensively involved with kernel testing. Our Scalable Test Platform and Patch Lifecycle Manager provide an open, publicly accessible workbench for the submission and testing of kernel patches.

The OSDL Database Test Suites are designed to provide the kind of complex database workload commonly found in data center applications. These are open source projects and continuing development and enhancement by the open source community is encouraged.

The 2.5 Kernel Stability Project brought together a variety of information on the current state of the "development" Linux kernel. You can find historical compilation metrics, current testing results, and information on OSDL's production use of the 2.5 kernel.

LWM: What was the goal with the creation of the Linux Legal Defense Fund? How has this fund been used so far?
Cohen: The OSDL Legal Defense Fund is designed to help defray legal expenses of Linux users involved in litigation with The SCO Group on issues that affect the Linux community and industry.

Really, we wanted to provide peace of mind to Linux end users and send the message that OSDL, along with its member companies and others throughout the Linux industry, will stand firm against legal threats raised by The SCO Group.

We have communicated with both Daimler Chrysler and AutoZone about the fund.

LWM: How do you foresee the future of Linux in the communication industry? What are the biggest challenges today facing the deployment of Linux on communication platforms?
Cohen: We believe that Linux is on its way to become the dominant platform in the communications industry.

Ironically, the biggest challenge in the communications business is just that - communication. The developers, network equipment providers, and distributions all have to communicate with each other regarding the requirements for next generation communications solutions.

LWM: Any final thoughts on Linux in the communication industry?
Cohen: CGL is a great example of major stakeholders cooperating and working together toward a common goal. In very little time, the members of CGL have created two specifications and are on track for a third.

Linux has the opportunity to become the basis for the next generation of communication software. We are seeing a revolution as customers and vendors move from the old "we build it all in-house" to a commercial-off-the-shelf world where everyone agrees to use common industry-proven building blocks for their infrastructure to save money and time.