Carrier Grade Linux: Adoption and Deployments
CGL is Real and Building Momentum...
The Open Source Development Lab (OSDL) released the latest version of the Carrier Grade Linux (CGL) Requirements Definition--version 3.1 on June 2, 2005. CGL 3.1 is the successor to CGL 2.0 and 1.1, the earliest versions of CGL which have been broadly adopted by the industry. At the time of writing, over 18 platform providers include CGL as part of their offerings and five Linux distributions are registered as implementing CGL 2.0 (with two more in progress). In addition, an increasing number of Carriers and Service Providers are deploying CGL based server nodes to offer communication services. In a previous article published in LinuxPlanet, I examined the state of Linux in telecom and reviewed the specifications in the CGL requirements document. In this article, I will provide an overview of CGL distributions, deployments and some of the challenges ahead.
Traditionally, communications and data service networks were built on proprietary platforms that had to meet very specific requirements in areas such as availability, reliability, performance, and service response time. Those proprietary systems were composed of highly-purposed hardware, operating system, middleware, and often included proprietary technologies and interfaces. Such proprietary approaches to system architecture fostered vendor lock-in, very served to limit design flexibility and freedom, and produced platforms that were and are very expensive to maintain and expand (see Figure 1).
Today, those same service providers and carriers are challenged to drive down costs while still maintaining carrier class characteristics for platforms to provide service and mission critical applications in an all-IP environment. They are in a position today where they must move away from specialized proprietary architectures, and towards commercial-off-the-shelf (COTS) approaches and building practices (Figure 2) for several reasons, driven by three key motivations:
- Faster time to market: Service providers and carriers need to be able to deliver new services based on common standardized platforms. They are in a constant race to deliver faster to the market. Building with propritary and specilized technologies that are offered by a very limited number of provides is one obstacle from this perspective.
- Reduce costs: Communications service providers need to reduce the design and operation costs by using commercial-off-the-shelf hardware and software components that are offered by multiple providers and are all compliant or registered towards standards of industry agreed specifications.
As a result, proprietary legacy systems no longer offer a viable approach. They are expensive to buy, maintain, and scale. As a result, the industry is moving away from specialized proprietary systems toward open platforms that are based on industry established standards and common practices. Figure 2 illustrates the first steps that were taken by the industry to move away from specialized architectures towards COTS practices and building blocks: the figure present a trend to move from a network element that was designed and built using proprietary components towards a network element which deploys standardized highly available hardware and Carrier Grade Linux.
As we move forward into the future (Figure 3), we will see the introduction of standard-based hardware manamagement, standard-based middleware and interfaces, and standardized protocols. The end result is a application enabling platform that encourages the rapid adoption of industry standards, promotes innovations, and promises lower costs.
Figure 4 illustrates a sample telecom platform, one subrack that consists of multiple network elements. Each network element runs a different type of application; all network elements have the same open architecture built around the concept of standardized building blocks and COTS hardware and software components.