I'm famous! (Or not.)

Okay ... I'm acknowledged! (Well, that is true. Even if it is only on page 116.)

The recent decision by the U.S. Department of Defense (DoD) to migrate from IPv4 to IPv6 in military networks underscores how policy can have a profound effect on network performance and even unintended consequences.

While to some IPv6 holds a promise of "increased security" and "support for wireless applications," these are very arguable points from an engineering perspective and, even if true, need to be seriously weighed against performance in environments with the unique link characteristics of military networks.

In "Internet Protocol Header Compression, Robust Header Compression, and Their Applicability in the Global Information Grid," E. Ertekin et al. show that moving from IPv4 to IPv6 increases the overhead due to packet headers by 50 percent. This will have serious performance impacts on capacity-constrained links, especially tactical radio links that need to keep the payload size small, resulting in a large percentage of the total payload dedicated to the header.

The authors present an excellent summary and comparison of current header compression schemes that can be used to mitigate this problem and present a good overview of the application of these techniques to the global information grid. Specifically, the article examines the use of two compression schemes developed by the Internet Engineering Task Force (IETF) and their applicability for military applications: Internet Protocol Header Compression (IPHC) and Robust Header Compression (ROHC). It is interesting to note that the DoD directive, while intending to standardize on IPv6 supposedly for increased robustness and interoperability, also trades that robustness for decreased link utilization and shifts the point of interoperability to the type of header compression a given system engineer chooses to implement.

Follow the first link ("Network-Centric Military Communications") to read the guest editorial. A subscription is required for access to the full text of the article.