Abstract

An important mechanism to build structured P2P systems are Distributed Hash Tables (DHT). DHTs are robust, scalable and have been researched for many years. DHTs usually offer the two fundamental operations put(key, value) and value = get(key), for retrieving and storing data. However, traditional DHT implementations have weak consistency capabilities for mutable data. An update usually overwrites old versions and leaving and joining nodes may cause updates to disappear. By slightly extending the traditional DHT operations get and put, this master thesis proposes a versioned DHT (vDHT) having stronger consistency properties for mutable data. Based on the proposed extensions and modifications, two putting strategies are described in order to update data in a DHT. The proposed putting strategies are analyzed in an evaluation by multiple simulations of a dynamical distributed environment, helping to compare the consistency of the vDHT approach to traditional DHTs. The evaluation showed that under worst churn conditions, one of the two proposed putting strategy achieves eventual consistency, the other one remains strictly consistent.