Content archiving systems today are facing growing data volumes and increasing document diversity. New requirements result from state and international laws as well as from user behavior shifting towards using full-text search to access information in all working environments. A project to develop and evaluate new approaches to the message retention problem is presented and discussed in regard to architecture and performance. An inverted index plays the central role in managing a large collection of e-mail documents. In an undertaking to attain optimal scalability, aspects and methods of index distribution in a duster environment are evaluated and search performance is measured. Strategies to improve scalability by using metadata-based index partitioning are presented and system design specifics are discussed. The prototype application is examined as an example implementation and research platform. Benchmark tests are defined and results for different setups are presented.