This article presents an overview of the initiatives at OSDL aiming to advance Linux adoption across multiple industries. We will cover the Carrier Grade Linux (CGL) initiative, the Data Center Linux (DCL) initiative, and the Desktop Linux (DTL) initiative.
The vision of CGL is to deliver next-generation and multimedia communications services using Linux-based open standard platforms for Carrier Grade infrastructure equipment.
The telecommunications industry is undergoing enormous change as equipment providers migrate from proprietary platform architectures to open software environments and commercial off-the-shelf (COTS) platform architectures. They see open software and COTS hardware as the means for rapidly deploying new voice and data services while reducing capital expenses and operating costs, which allows equipment providers to stay competitive and profitable.
Figure 1. From proprietary to standards-based platforms
Carrier Grade Linux (CGL) stands at the center of the move to open architectures. About three years ago, a group of industry representatives from platform vendors, Linux distribution suppliers, and network equipment providers set out to define how "Carrier Grade Linux" could enable environments with higher availability, serviceability, and scalability requirements. This led to the formation of the Open Source Development Lab (OSDL) CGL working group.
Since its formation, the working group has produced three versions of a specification to define these required capabilities. In response, Linux distribution suppliers now demonstrate that they can meet the emerging needs of telecommunications by registering (disclosing publicly) how their products address the requirements defined in the Carrier Grade Linux Requirements Definition document.
Today, the CGL working group has grown to include more than three dozen representatives from platform vendors, Linux distribution suppliers, network equipment providers, carriers, and development community members worldwide. This expanded group has released version 3.1 of the CGL requirements. A technology version of the document was released in early 2005, and a version that can be registered will be released in the second half of the year, for 2006 registration. For clarity and ease of use, the group has split the specification into seven topical documents: Availability, Clustering, Serviceability, Performance, Standards, Hardware, and Security.
Figure 2: CGL scope and architecture
As CGL capabilities become available in mainstream implementations and distributions, Linux not only becomes more attractive for telecommunications applications, but the entire Linux community also benefits from a highly available, scalable, high-performance, and manageable Linux environment.
A quick overview of the CGL specifications:
CGL 3.1 specifications are an upward-compatible superset of the CGL 2.0.2 specification. In 2003 and 2004, member companies produced communications products based on the 1.1 version of the CGL specifications. In the latter half of 2004, Linux distributors began to announce Linux offerings based on the CGL 2.0.2 specification. As the writing of this article, five Linux distributions have registered against the CGL 2.0: FSMLabs, Montavista, SuSE, TimeSys, and WindRiver. In addition, more than 20 platform providers offer CGL-based platforms and products. We expect that products based on the CGL 3.1 spec will begin to appear in 2006. We also expect a smooth transition for carriers and equipment providers as Linux distribution suppliers incorporate CGL 3.1 capabilities in 2006 and 2007.
As for beyond 3.1, the priorities of the CGL working group, as identified based on the market input and the feedback received from companies participating in the CGL initiative, are: real-time capabilities, testing CGL workloads, device driver hardening and availability, and Linux performance and scalability--in addition to further enhancements to security and manageability. In addition, much of the efforts in 2005 and 2006 will concentrate on promoting quality implementations of the CGL 3.1.
Figure 3: CGL initiative road map
Development is under way on many of the CGL capabilities that do not appear in mainline distributions. While the CGL requirements address Linux-based platforms in the communications industry, a high-availability, high-performance, scalable system is beneficial to the entire Linux user community.
Linux with Carrier Grade characteristics provides an essential building block that will let us build open communications platforms. CGL is a community effort: TEMs, NEPs, and carriers supply requirements; OSDL members gather requirements and create specifications; OSDL members, community, and Independent Software Vendors (ISVs) implement projects; distribution suppliers build and register CGL platforms; TEMs, NEPs, and carriers deploy. The initiative is based on cooperation between companies and individuals, and participation is open to everyone. Please consider this an invitation to get involved in this effort and contribute to the making of Linux as an alternative operating system for communications platforms.
Figure 4: CGL plans for 2005
This gives a brief introduction to the Data Center Linux initiative. It describes the goals, shows how OSDL achieves those goals though committees and working groups, and provides some examples of DCL-driven activities and challenges.
The OSDL Data Center Linux initiative was formed with the intent of promoting the adoption of Linux-based servers in the enterprise across its many tiers: edge, infrastructure, application, and database. It brings together interested parties to accelerate the availability of stable, fully featured, integrated, customer-available solutions that run on Linux. The initiative addresses both mid- and high-end multiprocessor servers as platforms for mission-critical enterprise applications and databases.
DCL has a three committees, composed of OSDL members and dedicated OSDL staff: Steering, Marketing Working Group (MWG), and Technical Working Group (TWG). The DCL Goals and Capabilities Document 1.1 documents the categories, marketing goals, and technical capabilities that MWG and TWG identify.
The scope of issues related to an Enterprise Data Center is extremely broad. Steering focuses the efforts of the initiative. For example, Steering directed that DCL identify the most important applications used in data centers today. From that, DCL can identify the marketing and technical inhibitors to running that application on Linux in the enterprise.
DCL has a marketing working group because many of the inhibitors to Linux adoption are not technical. In fact, the Linux operating system has improved so much that this is an increasingly common situation. MWG sets goals for nontechnical issues such as Linux awareness and confidence, global enterprise services and support, software availability for priority applications, training and education, care and feeding of the development community, stability, and total cost of ownership.
DCL's technical working group is responsible for identifying and removing the technical inhibitors to Linux adoption. When the initiative formed in 2002, the focus was on the Linux kernel. As the kernel has matured, TWG's scope has expanded to the whole software stack including and above the kernel.
TWG identified eight categories of data center concerns: scalability, performance, manageability, RAS (Reliability, Availability, and Serviceability), standards, security, clusters, and usability. To set the priorities of the effort, the group describes and maps the required capabilities in these categories according to the application workloads that marketing identifies (data warehousing for CRM, for example). For the highest-priority capabilities, the group identifies gaps in maturity and works to close those gaps.
DCL members want to be good open source citizens. Moving from a closed to an open source world requires a shift in culture for those making the transition. TWG assists those members' assimilation into the community. One DCL effort under way is to create an environment wherein more storage drivers exist exclusively in the mainline kernel. There is much good documentation on formatting style and coding rules for open source drivers. However, the knowledge about the right way to interact with the kernel community was missing and is under development.
As previously stated, the working groups identify and remove the technical and marketing inhibitors to Linux adoption. It is equally important to know that DCL is not another Linux distribution. To insure that solutions are available to customers, our approach is always to drive kernel-based solutions into the mainline kernel.org kernel. Any open layered solutions go through community recognized forums.
A challenge for DCL's TWG, and indeed for all the initiatives, is to map the capabilities needed for:
To overcome these challenges, OSDL created special interest groups (SIGs) in high-priority subject areas (for example, storage networking, hot plugging, clusters, and security), which our members use as cross-initiative technical forums to identify the use cases and gaps (coding, testing, and documentation) needed to provide enterprise-worthy solutions on Linux. These forums are completely open so that we can include nonmember maintainer participation. The intent is never for SIGs to replace active community developer mailing lists. If the existing community can deal with an issue, there is usually no need for activity on the part of a SIG. SIG activities are public on their home pages.
One example of a SIG activity is that of the Storage Networking SIG, which has a focus area of NFS Version 4. The SIG determined that the development community was progressing with no need for intervention, except that no one could answer the question "What testing is necessary to make NFS V4 ready for customers, and who will doing the testing?" The SIG sponsored an effort to create a prioritized NFS V4 test matrix to do the following:
This information helps the community know what testing is highest priority to more effectively assign test resources and not duplicate efforts. It also means that it is possible to identify the testing gaps and raise them to DCL Steering for action. The SIG gathers information from the NFS V4 developers on their own forums. Developers do not have to join the SIG, but those who wish to are certainly welcome.
The OSDL Desktop Linux Initiative (DTL) formed with the intent of promoting the adoption of Linux desktop systems in the enterprise. It does not specifically address Linux desktops in any other area, but the group does of course have a strong interest in what happens in those areas.
DTL has a workgroup composed of OSDL members and dedicated OSDL staff. The initial work involved determining a set of usage models that accurately represent the majority of desktop uses over a broad range of enterprise use. The group eventually decided upon five usage models:
The intent of the group is to create a list of the capabilities that a desktop system must have to successfully address each of the usage models. Once the group understands and clearly documents the required capabilities, it then becomes possible to identify key inhibitors that are preventing successful adoption, as well as specific technologies that either are not present or have some deficiencies when applied to enterprise environments. Working with Linux distributors and existing open source development communities, and, if necessary, creating new development communities by way of OSDL SIGs, the group hopes to accelerate Linux development in the specific areas that will facilitate its adoption on the enterprise desktop.
After deliberation, the group decided that attempting to address all five of these usage models would involve a scope too broad for the team to tackle realistically. The Advanced Workstation usage model had by far the most complexity and broadest scope, but it actually reflected only a small percentage of desktop use within most enterprises. The initial work thus focused only upon the first four usage models.
It soon became obvious that one of the most important limiting factors for the enterprise was the availability of commercial software packages. After some investigation, the group realized:
One of the prime areas of focus has become enabling ISVs on Linux, since this addresses not only fundamental issues, but also issues that are of more general interest.
In February 2005, DTL produced a document that was not meant to be complete but simply a snapshot document of the work to date. This document, DTL 1.0 Capabilities, is available on the OSDL web site.
In March, DTL held a strategy meeting in Boca Raton, Florida, to review progress and to further refine its focus. Work on that continues.
OSDL Developer Resources
Lynn de la Torre is a member of OSDL and coordinates the activities of the DCL Working Group.
Ibrahim Haddad is the Director of Technology for the Software Operations Group (Home & Network Mobility Business Unit) at Motorola Inc.
Philip Peake is a member of OSDL and coordinates the Desktop Linux Working Group.
John Cherry is the Roadmap Coordinator for the Carrier Grade Linux initiative at OSDL.
Mary Edie Meredith is a member of the OSDL engineering department and is Roadmap Coordinator for the Data Center Linux initiative.
Return to the Linux DevCenter.
Copyright © 2009 O'Reilly Media, Inc.