Distributed Multimedia Synchronization Based on Fuzzy Causal Relations

In this dissertation, the research is focused on the field of distributed multimedia systems (DMS). One of the main problems in DMS is the data synchronization. Synchronization is concerned with the preservation of temporal dependencies among the application data from the time of generation to the time of presentation. The synchronization problem can be characterized as an event ordering problem. Event ordering addresses the problem of establishing a certain order among the events that occur in a distributed system (DS) according to some particular criteria. The types of event orderings used in a DS are: no order, FIFO, causal, ∆-causal, total, and causal-total. They mainly differ in the degree of asynchronous execution allowed. One of the most important orderings is the causal order (CO), which is based on Lamport's happened-before relation. It establishes that the events must be seen in the cause-effect order as they occur in the system. However, for certain applications, for example multimedia synchronization, where some degradation of the system is allowed, ensuring the CO based on Lamport's relation is rigid and negatively affects the performance of the system. In this dissertation a new ordering for DS is introduced in order to achieve a more asynchronous execution than the CO. This new ordering is called Fuzzy Causal Order (FCO). In addition, the Fuzzy Causal Relation (FCR) and the Fuzzy Causal Consistency (FCC) are defined. The FCR establishes logical dependencies based on the precedence of events and by considering some kind of " distance " between their occurrences. With the notion of distance it was possible to establish a cause-effect measure between two events a and b that indicates " how long ago " an event a happened before an event b. Through the FCC, it was possible to determine " how good " the performance of the system is at a given moment. The usefulness of the FCO, FCR and FCC is shown by applying them to the concrete problem of intermedia synchronization in DMS. In order to overcome the synchronization problem based on these concepts, a distributed multimedia model and a synchronization algorithm were designed. In addition, a fuzzy control system to adjust the delivery time of the messages and to determine if a selective message discard must be carried out was designed.