Analysis of Binary XML Suitability for NATO Tactical Messaging

The ability to efficiently transfer information among tactical systems is essential for network-centric operations. However, maintaining interoperability among heterogeneous networks and applications is a challenging issue, especially for large enterprises such as the US Department of Defense and NATO. Each of these organizations maintain extensive communication networks of tactical systems that process and manage all types of data. Additional complexity is added when considering that many systems are built with a variety of proprietary or legacy data formats. Establishing and maintaining interoperability is difficult. Using XML, many interoperability issues can now be successfully addressed. XML provides a self-describing way to effectively structure information that can be applied to compose diverse tactical communications. However, XML is inefficient for network transmission since it uses a text-based format which can consume more memory (and thus more bandwidth) than binary equivalents. In addition, parsing text-based documents is slow and computationally expensive. One potential solution is to use GZIP to reduce the file size before transmission. Unfortunately, this solution has limitations since it often provides suboptimal compression and also requires additional processing time when extracting data. Recent standardization efforts have identified promising new encodings for XML that use binary representations to reduce parsing time, memory size, and bandwidth requirements. This thesis surveys conversion of NATO tactical data link information into an XML format for distribution to command and control centers. General benefits and tradeoffs are then considered for applying binary XML encoding to that data. This thesis also examines work done by the World Wide Web Consortium in examining common use cases and developing the requirements needed for a binary XML encoding. The performance of two specific implementations, XML Schema based Binary Compression (XSBC) and Fast Infoset (FI), are compared with GZIP. XML files of varying sizes are encoded in binary form, then compression ratios and parsing times are compared and analyzed. Initial results are excellent and further work is recommended.