The CYCLOPS Platform: geospatial services implementation Chania 5th May 2008 CYCLOPS Second Training Workshop The CYCLOPS Platform: geospatial services implementation Valerio Angelini1 Paolo Mazzetti1 Stefano Nativi1 Paolo Fiorucci2 Marco Verlato3 1CNR-IMAA 2CIMA Foundation 3INFN angelini@imaa.cnr.it

Outline The Civil Protections requirements The CYCLOPS architecture The gridded OGC Web Coverage Server Use case: the RISICO forest fires assessment application Porting RISICO in the CYCLOPS Grid infrastructure throug OGC Standard Web Services CYCLOPS, as you probaby know, is a SSA founded in FP6 which aims to study a cyber infrastructure to connect the various European Civil Protections Our work started from the analysis of the Civil Protection requirements From these requirements we derived our CYCLOPS architecture Recently we chosen a particular operative use case, the RISICO forest fires assessment application, and we ported it to the Cyclops Grid Infrastructure

From CP/GMES needs Civil Protection and GMES applications/systems have specific requirements: to share geospatial information that has complex characteristics: Huge amounts of remotely-sensed observations, which are multidimensional and frequently updated to access infrastructure, run models and search information in a real-time (RT) or near-real-time (NRT) way, privileging time of response instead of accuracy have access to computational power for working out simulation tasks or tasks for critical missions To implement the strict data policy and the security requirements typical of dual systems (civil/military); To formalise the knowledge required to analyse data and provide decision-makers with effective information They need At least

To GRID (Why to use Grid Platform for GMES applications?) Provide physical resources to support heavy application workloads Virtual Organization management Provide a security infrastructure Provide a quality of service agreement & monitoring infrastructure We have chosen the grid tecnology to be the base of our architecture The grid technology

Communities Interoperability Semantic Level GMES & CP Communities Strategic Applications Geospatial Community Geospatial Information services Goal is to bring together these two communities: BLA BLA Resources Provision GRID Community

CYCLOPS Architectural Framework Appl. Platform Real Time and Near Real Time Applications for Civil Protection (Data integration, high-performance computing and distributed environment for simulations) CYCLOPS Platform Security Infrastructure Business logic Services Presentation and Fruition Services Interoperability Platform CYCLOPS Infrastructure Spatial Data Infrastructure Services Environmental Monitoring Resource Infrastructure Geospatial Resources Services Advanced Grid Services I will show you the architecture we developed: GRID Platform (EGEE) Processing Systems Infrastructure Data Systems Sensor Element Services

Traditional WCS Data Request Response WCS interface Subsetting Resampling Data Encoding Request Response WCS interface

GRID WCS CE CE CE Workload Manager Result URL WCS Interface SE SE Subsetting Data Resampling CE CE CE R D S Data Encoding EGEE Grid Interface EGEE Grid Interface EGEE Grid Interface WCS interface Workload Manager EGEE Grid Interface Result R D S Result URL WCS Interface Async. response Request

GRID WCS : Technologies SE SE EGEE GRID Computing Layer (Java) CE CE CE R D S EGEE Grid Interface EGEE Grid Interface EGEE Grid Interface Java Serialization Workload Manager JDL EGEE Grid Interface Result R D Application Layer (Java) S HTTP/SOAP WCS Interface Async. response WEB Request

RISICO use case: Forest fires risk assesment RISICO is a forest fire risk forecasting system, operative for the Italian Civil Protection (DPC) since august 2003 It is developed by the CIMA Foundation for the DPC. Such a system is made of two main models: the fuel moisture model the potential fire spread model the fuel moisture model representing the dynamic behaviour of the distribution throughout the territory of the variable expressing the water contents of the vegetation; the potential fire spread model evaluate the physical characteristics that a fire could assume, on the basis of the variables that locally determine the possibility of a successful ignition and fire propagation.

RISICO current implementation RISICO current implementation runs on a single node on the CIMA premises It accesses LAM and Earth Observation input data through heterogeneous protocol and interfaces Heterogeneous interfaces/protocols Processing Server Running RISICO on a time base LAM data EO data model output data GUI EO data

RISICO execution on Grid 2 CYCLOPS WPS GUI GRID CYCLOPS WEB LAM data EO data Old Input Files Computing Elements WCS CYCLOPS Model Input files 1 GRID The old in put files are stored in the Grid storage elements The WCS dynamically generates the model input files, subsetting and/or resampling the origin files stored on the SE Storage Elements

Gridded WPS Activity diagram Through standard Access services Parallelized on grid Can handle response time (CP Requirement) All begins with an WPS Request Every process defines a “job splitting” strategy

RISICO advanced data access RISICO GUI CYCLOPS WCS WFS WPS request WCS GetCapabilities CS-W GetRecords WEB CS-W WCS requests CYCLOPS WPS Making use of standard web services CYCLOPS WCS GRID

RISICO improvements Flexibility: The geospatial services layer allow to: Integrate new and heterogeneous input data. Dynamically choose the spatial location and resolution of the run. Be interoperable with other “standard-based” infrastructures. Scalability. The grid infrastructure provides high processing and storage capabilities on-demand, allowing to: Improve output data resolution. Widen the covered area. Improve the response time. (Futures) Improve model complexity. The introduction of the geospatial services layer allowed

Architecture main achievements CP Requirement Strategy sharing geospatial information OGC Web Services strict data policy Grid security layer / VO management real-time or near-real-time Advanced grid services computational power for working out simulation task Grid computing capabilities Large and fast storage Grid storage capabilities

Progress of Work The RISICO prototype that makes use of the CYCLOPS platform is in the deploying phase The prototype will be presented in the next IBERGRID ‘08 in Porto (12-14 March 2008)

Thank you for your attention ! I’ll be happy to answer you questions

Back-up slides

Web Site Updated and revamped Contacts Corrado De Rosa (DPC) Valerio Angelini (IMAA) http://www.cyclops-project.eu

Forest Fire Risk Management application A system for wildland fire hazard mapping and resource management Vegetation modelling, along with meteorological data and topography, allow simulating moisture dynamics of different kinds of dead and live fuels over a given area A semi-physical fire propagation model gives a quantitative evolution of the hazardousness

WEB CYCLOPS CYCLOPS CYCLOPS WPS Basic functioning GRID RISICO GUI WPS request WEB WCS request CYCLOPS WPS CYCLOPS WCS GRID

GUI Preview