Scalable indexing and query processing of big spatio-temporal data
Κλιμακώσιμη ευρετηρίαση και επεξεργασία ερωτημάτων για μεγάλα χωρο-χρονικά δεδομένα
Keywords
Indexing ; Querying ; Data processing ; Big data ; Spatio-temporal data ; NoSQL ; Spatial joins ; Weather data ; Column-oriented file formatAbstract
Our era is characterised as the “Big Data Era” where the volume of generated data grows exponentially. Much of the generated data capture the information of events and phenomena that unfold in both space and time. This kind of data is also known as spatio-temporal data, found in several domains such as urban planning, transportation logistics, epidemiology and environmental monitoring. Analysing such data in massive scale, can unveil patterns and trends from which valuable knowledge can be extracted. Even though traditional data management approaches have been studied extensively, it is still challenging to support efficient operations in highly scalable environments. The use of new data management systems and algorithms in distributed environments is necessitated to meet the requirements of big data volumes. Towards this direction, this dissertation focuses on the data management subject areas of i) storage ii) indexing iii) querying and iv) processing of big spatio-temporal data. The proposed methods and algorithms improve the application of widely-used stores and frameworks in the big data ecosystem for spatio-temporal data, which they do not inherently support. The solutions can scale out efficiently, making them suitable for data-intensive cases.