Welcome to Hermes Moving Object Database Homepage
Moving Object Database (MOD) engines enable us to process, manage and analyze mobility data. Hermes provides MOD functionality to OpenGIS-compatible state-of-the-art Object-Relational DBMS. In general, the big picture of Hermes aiming for the efficient management and exploration of mobility data is illustrated in the following figure.
According to Fig. 1, the telecommunications network provides raw unprocessed data (e.g. GPS recordings), which, in turn, is processed and trans-formed in mobility data (e.g. trajectories of moving objects) stored in MOD systems (in their detailed form) and Data Warehouse systems (in aggregations over detailed information, respectively). The stored information can be queried as well as provide input for advanced analysis, such as multi-dimensional (OLAP) analysis and data mining, the output of which is appropriately visualized to end users.
Key questions arise from the above architecture:
– How to reconstruct trajectories from raw GPS logs? How to store trajectories in a database system?
– What kind of analysis is suitable for mobility data (in particular, trajectories of moving objects)? How does infrastructure (e.g. road network) affect this analysis?
– Which patterns / models can be extracted from them (for example, clusters, frequent patterns, anomalies / outliers, etc.)? How to compute such pat-terns / models efficiently?
– How to protect privacy – user anonymity? What is the tradeoff between privacy protection and quality of analysis?
– Which are the semantics hidden into the mobility data we gather? How can we make them “first-class citizens”?
– What if mobility-aware applications become so widespread, so that extremely large volumes of data bomb servers? What is the position of mobility data management and exploration in the big data era?
The above give only a few hints of what is studied in our recent book, and is exactly the challenge accepted by Hermes. Currently, Hermes comes in two implementations; the first operates on top of Oracle DBMS (coined Hermes@Oracle) and the second operates on top of PostgreSQL (coined Hermes@Postgres). Although ultimately the two versions support more or less the same functionality based on a similar model, there are few differences. In the reminder of this page, we present Hermes implying the version on top of Oracle. Abstractly, Hermes defines a trajectory data type and a collection of operations, which is further enhanced by special trajectory preserving access methods. Fig. 2 illustrates Hermes@Oracle architectural framework.
According to Fig. 2, Hermes resides at the ORDBMS tier. In detail, the Oracle ORDBMS Server enhanced with trajectory data storage and query capabilities serves as the infrastructure for mobility data management. In order to implement such a framework in the form of a data cartridge, Her-mes exploits a set of standard data types together with (stationary) spatial data types offered by Oracle Spatial and appropriate temporal types. Em-bedding this functionality in the database query language of the underlying ORDBMS, one obtains an expressive and easy way to manipulate moving objects. Accessing Hermes’ API can be done exactly the same ways as any native component of the underlying ORDBMS.
If you have any publications resulting from this system, please cite the following paper(s):
- N.Pelekis, E.Frentzos, N.Giatrakos and Y.Theodoridis. Mobility Data Management and Exploration. Springer, New York, 2014, in press. ISBN 978-1-4939-0391-7.
- N.Pelekis, E.Frentzos, N.Giatrakos and Y.Theodoridis. “HERMES: Aggregative LBS via a Trajectory DB Engine”, In the Proceedings of the ACM SIGMOD Conference, Vancouver, 2008.
- N. Pelekis, E. Frentzos, N. Giatrakos, and Y. Theodoridis. “HERMES: A Trajectory DB Engine for Mobility-Centric Applications”. International Journal of Knowledge-based Organizations, (IJKBO), 4(1), 2014.