Open Geospatial Cloud Service for Research to Support Ubiquitous Data Driven Discovery
ISBN 978-85-88783-11-9
Authors
1Sarjakoski, T.
1FINNISH GEDETIC INSTITUTE Email: tapani.sarjakoski@fgi.fi
Abstract
During the last ten years, tremendous developments have taken place, concerning users of geospatial information. Initiatives on spatial data infrastructures have given promises of efficient, distributed management, easy discovery, and usage of data access services as well as processing services. Cloud computing platforms have emerged to offer extreme flexibility for storage and processing of geospatial information. New sensors, either mobile on the ground, on the sky or in the space, have started to deliver continuous streams of Earth related data. These datasets are so vast and varying that they deserve to be called Big Geospatial Data. Motivated through many of the developments presented above, especially the communities producing and maintaining massive databases of satellite imagery have made a conclusion: because of the volume of the data, it is necessary to provide storage and processing services to users from centralized cloud services. In this paper we want to elaborate that idea from a broader perspective and pose a question: would it be a right time and a right direction now to establish national cloud-based service centres providing geospatial storage and processing services especially for research and education? In order to answer the question above, we will present a thorough review of existing geospatial cloud services. We will also analyse user requirements in detail. We will also make a rationalized suggestion on how an Open Geospatial Cloud Service for Research should be established at national level. As a framework for the review and analysis we use ISO/IEC standard 10746-1, Open Distributed Processing (ODP). The ODP defines five different viewpoints from which a distributed system can be described: 1) enterprise viewpoint, 2) information viewpoint, 3) computational viewpoint, 4) engineering viewpoint and 5) technology viewpoint. A spatial data infrastructure is commonly understood as a data infrastructure implementing a framework of geographic data, metadata, users and tools that are interactively connected in order to use spatial data in an efficient and flexible way. From a users point of view, a cloud-based service centre for geospatial data can fulfil most of the criteria of spatial data infrastructures. The essential difference we take in this presentation is, however, that we assume the service to be provided from a service point having a single physical location. In other words, the hardware for the storage and computations are truly centralised. As observed above, this is a prerequisite for efficient computing in advanced geospatial analysis utilizing big geospatial datasets. Another difference that we want to emphasize here is that the dataset managed in this centralized service should be mutually harmonised so that data interoperability is at workable level. This applies not only for the technical aspect such as formats, encodings and coordinate systems but also for the content related aspects of the datasets, including issues such as semantics and level of detail of the information in the data sets. It is self evident that many of the datasets to be incorporated into the envisioned cloud service have to be imported from external origins using automatic import mechanisms. The goal for setting up a national Open Geospatial Cloud Service is to serve researchers in many fields and to provide them with a good environment for geospatial analysis. Integration of many harmonized data sets into a single environment can guarantee a plug-and-play compatibility among the data sets. Good interactive analysis and visualisation tools are necessary for explorative data driven discovery on Earth-related phenomena. Functionalities to manage workflows are important for a single user as well as for collaborative work among larger research communities, to be able to share and accumulate knowledge on geospatial analysis.
Keywords
cloud service; big data; geocomputing