SensorML for Grid Sensor Networks Giovanni Aloisio †† , Dario Conte †,†† , Cosimo Elefante † , Italo Epicoco †† , Gian Paolo Marra †,†† , Giangiuseppe Mastrantonio † and Gianvito Quarta †,†† †) Institute of Atmospheric Sciences and Climate of the Italian National Research Council via per Monteroni, 73100 Lecce – Italy ††) Center for Advanced Computational Technologies/ISUFI, University of Lecce, at NNL/INFM&CNR Lecce, Italy {d.conte, c.elefante, gp.marra, g.mastrantonio, g.quarta}@ isac.cnr.it, {giovanni.aloisio, dario.conte, gianpaolo.marra, gianvito.quarta }@unile.it Abstract - This paper describes an approach based on Globus toolkit for developing grid sensor networks. The key aspect is also represented on the use of a novel information service based on a relational data model, namely iGrid developed within the European GridLab project. iGrid is used to integrate sensor networks in Grid environments by means of the design of an information structure based on Sensor Modeling Language (SensorML). A case study is also presented in order to provide some relevant details about our approach and to verify the concrete applicability of the proposed methodology. Keywords: Grid Sensor Networks, Grid Computing Environment, SensorML. 1 Introduction A sensor network [1] is a computer network of many spatially distributed devices using sensors to monitor conditions at different locations. Such distributed devices are typically used to monitor temperature, sound, vibrations, pressure, motion, pollutants and so on. Usually these devices are small and inexpensive, so that they can be produced and deployed in large numbers therefore, their resources in terms of energy, memory, computational speed and bandwidth are severely constrained. Due to the increased use of this technique in several applicative domains, the requirements, in terms of processing, data gathering, data storage and mining were rapidly increased. At the same time, Grid computing [2,3] has evolved as standard-based approach for heterogeneous and geographically spread resources integration and access. In fact, computational Grids are spreading as technology that allows the sharing and utilizing of computing resources, software, knowledge, scientific tools and so on, in an efficient and coordinated manner. Several developments in Grid computing have focused on compute and data grids as suitable solutions for putting together computational power and provides a seamless access to very large amounts of storage resources. Another domain in which the grid paradigm is beginning to be used is the sensor networks. In fact, sensor grids extend the grid approach to the sharing of sensor resources in wired and wireless networks. As demonstrated by several works and mentioned in [4,5,6], sensor grids are suitable for improving the sensor network with regard to the following issues: • sensors can be different in terms of providing information and its data format; • sensors can be owned by different organizations which would like to share the sensors in a controlled manner; • sensors can continually produce a very large amount of data; • particular prominent information can be found if a powerful tool to query the sources is provided. Grid approach promises to efficiently solve these issues providing a suitable technology through which large amounts of data can be collected, stored and processed, using a data grid from one side, and a computational grid from the other. Moreover, a very prominent feature of grid approach is: several Virtual Organizations (VOs) [7] can by easily defined, in order to provide fine grained access control to geographically spread sensors. In this paper, we propose an approach based on iGrid information service to integrate sensors in grid environments. In fact, sensors can be considered as grid resources that provide specific information about some phenomena. The sensors are characterized by a strong heterogeneity, a variable state, are often geographically spread and can be shared among many organizations. In this paper we also describe Ligris, a case study in which our approach used, to put a SODAR (SOnic Detection And Ranging) sensor into a grid environment. In the next section, we describe some relevant related