An insider´s guide to logic in telecommunications data

Most of the functionality of modern telecommunication systems has migrated to software, and much of that software consists of code for manipulating data: data describing either subscriber features or equipment configuration. The data is generally hierarchically structured but extremely complex - it consists of hundreds or even thousands of "classes" (roughly speaking, distinct labels on a tree), with the same information occurring in different formats across many classes in a hierarchy. The data is also required to satisfy a gigantic number of integrity constraints, vital to the correct routing of phone calls. The challenge is to create tools for populating this data, updating the data, and propagating these updates to derived data; all without violating these integrity constraints. The aim of this article is to give an overview of recent projects at Bell Labs that tackle this problem. Declarative languages play a key role. Indeed, we can think of a solution as consisting of languages for: the integrity constraints, the large-scale changes to the data, and the relationships between base data and derived data. In this article the author starts by explaining the languages used, and the key algorithmic ideas behind the tools. Generally these tools can all be seen as "implementations" of closure properties of definable tree and tree-structured data languages. The author then discusses briefly some of the pragmatic issues in getting these systems used in industrial projects.