Improve the efficiency of your information distribution network! This paper presents a method for updating read-only copies (data replicas) in a distributed information network. It addresses scenarios where immediate synchronization with the source data isn't required, allowing an arbitrary mapping from source records to replica records. This method is designed to be fail-safe, maximize workstation autonomy, and suit networks with slow or unreliable communications. The algorithm centers on query manipulation, using short encodings to represent queries and recording the portion of the source database used when generating a response. This information then identifies obsolete replicas during source data updates. The workstation process deletes each obsolete replica and replaces it by an up-to-date version. The proposed method ensures atomic update transactions at each workstation. This research offers valuable guidance for designing and managing distributed information networks, especially in environments with limited or unreliable communications, enhancing data consistency and system resilience.
Appearing in ACM Transactions on Database Systems, this paper directly aligns with the journal's focus on managing and optimizing database systems. The presentation of a method for updating data replicas in distributed information services contributes to the journal's collection of techniques for ensuring data consistency, reliability, and efficiency in distributed database environments, appealing to researchers and practitioners in database management.