Optimisation and evaluation on data services based queries
LOPEZ ENRIQUEZ Carlos Manuel
under the direction of Christine COLLET and Genoveva VARGAS SOLAR
Funding: ANR OPTIMACS
Today there is a need to query large amounts of widely distributed data that are produced by different sources of data viewed as services (eg, sensors, devices, networks). Optimization of query evaluation based on services is a significant challenge to get the right information at the right place at the right time in an efficient way. This means to be able to evaluate classic queries, mobile and continuous queries by coordinating services that are nomad and static data providers producing data continuously or in batch. The theme of the thesis deals with multi-objective optimization of queries implemented by services coordinations (ie, query workflow). The idea is to provide a tool for enabling efficient access to data services, capable of optimizing query evaluation implemented by a services coodination. The objective is to combine services, and data streams with query evaluation while considering aspects of quality of service associated with the use of services.
QoS-aware Event Processing Systems for Wireless Sensor Networks
GUTIERREZ MLOT Esteban
under the direction of Christine COLLET, Paolo PAGANO (CNIT, U. Santa Ana, Pisa, Italy), Genoveva VARGAS-SOLAR
Funding: CNIT Lab, U. Santa Anta, Pisa, Italy
Event-based systems (EBS) can offer anonymous and asynchronous communication (producers/servers and receivers/client do not know the identity of each other) to make interoperability and collaboration between autonomous and heterogeneous components easier. EBS can observe, convey, filter, aggregate, compose and notify events. Efficiency, scalability and reliability are among the most critical aspects to be considered in this sense. They have to be guaranteed by means of appropriate QoS management and event composition techniques. The correct definition of the appropriate QoS metrics and the consequent implementation of prioritization and differentiation techniques helps to balance the load on resources and to optimize data routing. Through event composition, the unnecessary transmission of data is avoided thanks to data aggregation and filtering. Wireless Sensor Networks play a crucial role in data acquisition and event management. The new generation of data acquisition systems shall be able to get data from heterogeneous sensors packed in embedded devices and manipulate such data in order to produce semantically rich events conveying data related to different contexts and priorities.
The objective of the doctoral thesis is to define new event composition and QoS management techniques especially targeted to Wireless Sensor Networks. In order to support the heterogeneity of network nodes and sensor platforms, the techniques will be integrated in an existing middleware interoperable with widely accepted network standards, such as IPv6.