Keeping Up with Carrier Grade

by Ibrahim Haddad

Welcome to Ottawa!

A few weeks ago, I had the great opportunity to present at the Ottawa Linux Symposium (OLS). For those unfamiliar with OLS, it is a yearly Linux technology event, begun in 1999, during which software designers and developers working on the Linux kernel get together to present their activities, to discuss features and projects and, of course, to have a beer or two. This year's event was among my favorites. It was filled with rich sessions and lively discussions, and many topics that were covered. Over 600 attendees turned out for this year's event, making it the place to be.

Shameless Self-Promotion

As a presenter, I chose to cover one of the current technology trends we are witnessing in the telecom sector. Open and standardized platforms that use software and hardware component building blocks are, where suitable, complementing our existing telecom solutions. The Linux kernel as an operating system is a good example. The presentation was based on a submitted paper titled "Towards Linux-based Open Telecommunication Platforms".

In my opinion, I think this trend is good news for the Linux community. After all, the Linux kernel can be deployed as an operating system, running side-by-side with other operating systems. Some features and mechanisms still are needed in Linux, however, to support this transition.

I also wanted to present to the community some of these enhancements that are essential in a telecom environment needed by carrier grade platforms that have strict requirements in areas of performance, soft real time, availability, security and reliability. To be specific, the enhancements I presented are contributions from Ericsson, which open-sourced these technologies to enrich the Linux kernel with features needed on server nodes operating in mission-critical environments.

The contributions fall in the areas of cluster communication protocol (TIPC) and cluster security (DigSig). They also cover networking to improve route lookup and to support a multiple forwarding information basis, as well as a low-level kernel mechanism (AEM) for improved reliability and soft real-time performance.

The presentation went well, and I have received a lot of questions since then, mostly regarding the projects and the implementations. In addition, that same day, the Ericsson Linux Web site, received a lot of traffic, thanks to the availability of a wireless connection in the conference venue.

Questions, Anyone?

During the first ten minutes of the presentation, I felt that there wasn't much interest in the technology trend. This can be interpreted as the audience is a special audience that lacks interest in technology trends, which also can be considered to be an indicator of their own priorities. However, when I started to talk about technical contributions, open-source projects--their status and plans--people started to ask a lot of questions. What was most interesting is that people attending the presentation were answering many of the questions asked by others. I took this as a good sign that these contributions are being noticed--some people already are aware of them and know enough about them to answer questions.

Why am I talking about this? Well, my point is, within this community, source code is king.

Carrier Grade Linux

When talking about Linux in telecom, people often think of the ongoing activity in the Open Source Development Labs: Carrier Grade Linux (CGL). OSDL has a working group, CGL WG, that was established in January 2002 with the goal of accelerating the adoption of Linux in the telecommunication sector. Since then, the working group has produced the CGL 1.0, 1.1, 2.0 and, most recently, the 3.0 requirements definition document, the later being a public draft. A lot of work has been invested in this working group by many devoted and hard-working people from different companies, all of whom believe that Linux is the operating system that could be running on many pieces of telecom platforms.

The CGL WG at OSDL also has a sub-group called the PoC WG, Proof-of-Concept. In my opinion, the PoC WG is an important aspect of the CGL work, because it is responsible for driving the implementation and integration of carrier grade enhancements to Linux, as identified and prioritized by the requirements document. The PoC WG also is responsible for establishing and leading an open-source umbrella project to coordinate implementation and integration activities for carrier grade Linux enhancements. That is a challenging and interesting mission, and having enough resources in this WG is a necessary prerequisite for the CGL technology to be adopted. However, resources always are an issue.

OSDL CGL and the Community

On October 16, 2002, I had the opportunity to meet Greg Kroah-Hartman when he was invited to attend one of the CGL WG meetings. Greg was asked to make a presentation on how to work with the Open Source community. Greg gave a good presentation and had a lot of examples of how to do things the right way and how to do things the other way. He also explained the best approaches to work with the community. It was an session, especially because some of the attendees had little prior interaction with the kernel development community.

I encourage the Open Source community to take a close look at what OSDL is doing with regards to CGL. The CGL specifications, the source code contributions by member companies and the work of the PoC group to implement the specification represent a valuable contribution to Linux.


The CGL WG has a challenging mission. A lot of work has been done and much more remains to be done, especially in such areas as security and kernel stability; these are areas in which the PoC WG can contribute, assuming it has enough resources. With the ultimate goal being to improve Linux and make it succeed in the telecom space, everyone's involvement, cooperation and commitment is needed.

Bon Courage, and for all those participants at OLS, happy hacking, keep up the good work and see you next year!


Many thanks go to Andrew Hutton and the OLS team for organizing such a great event and to my colleagues who reviewed this article.

Ibrahim Haddad works for the Ericsson Research branch in Montreal, Canada. He also serves as contributing editor to Linux Journal. Ibrahim co-authored with Richard Peterson Red Hat Linux Pocket Administrator and Red Hat Enterprise and Fedora Edition: The Complete Reference (DVD Edition), both published by McGraw-Hill/Osborne. He currently is a Dr. Sc. Candidate at Concordia University.

Load Disqus comments