Presentation is loading. Please wait.

Presentation is loading. Please wait.

Swiss Experiment EPFL-LSIR Report Hoyoung Jeung SwissEx Annual Meeting, Zurich 15 th June.

Similar presentations


Presentation on theme: "Swiss Experiment EPFL-LSIR Report Hoyoung Jeung SwissEx Annual Meeting, Zurich 15 th June."— Presentation transcript:

1 Swiss Experiment EPFL-LSIR Report Hoyoung Jeung SwissEx Annual Meeting, Zurich 15 th June

2 Deliverable Overview

3 Framework Overview captured queried processed

4 Data Integration  Requires close collaboration with end-users –BIG efforts to find appropriate solutions to very specific needs –Continuous migrations of data along with hardware evolution and addition of new types of sensors.  in 2010 –SwissEx general (hosted by SLF) – http://montblanc.slf.ch –LTE – http://ltepc.epfl.ch –APUNCH – http://ifu-apunch.ethz.ch:22001/ –PermaSense – http://data.permasense.ch  Continuing in 2011 –CCES general portal - http://montblanc.slf.ch:22001 –CCES public data - http://montblanc.slf.ch:22002/ –Radar images and data from APUNCH: http:// ifu-apunch.ethz.ch:22001 –Disdrometer and radar data from Environmental Remote Sensing Laborator y: http://ltepc3.epfl.ch:22001

5 Data Repository/Mirroring Services  In addition to the aforementioned instances of GSN, mirror some of the data using additional machines at EPFL. –Near real-time data duplication. –Multipurpose: backup, query load distribution. –Accessible only for authorized users  LSIR repository –Mainly data from SLF, backup purpose  Lafouly repository –Data from a catchment used by the EU project Hydrosys in Lafouly region. Mainly used for GEOtop model simulations.  PlanetData repository – Mirroring Wannengrat public data at SLF for query load distribution.

6 Spatial Data Support  Extending GSN for handling geographical data  Support spatially-enabled processing –Data types: point, polyline, polygon,… –Spatial queries: e.g., average wind speed in Lausanne yesterday?  Two implementations –Java topology suite (JTS) integration: used for any DB –PostGIS integration Only for PostgreSQL DB Support spatial index  Output in GML format

7 Sensor Data Search  Processing in a federated network –Wiki: central point, easy to use GUI based on Google Maps interface –GSN instances: distributed, very fast data visualization  Ontology-based –“temp=temperature=air temp” –W3C SSN + NASA SWEET –On-going work  Plan to scale up –Search beyond swissex data

8 User-Interactive Data Cleaning  Model-based anomaly detection –User can select model/parameters, parametric/non-parametric original data stream ↓ approximation using user-selected models ↓ detecting anomalies ↓ user confirmation: anomaly is an actual error?

9 New Wrapper Developments  Grid wrapper –Reads ESRI grids (e.g., digital elevation models) from files and stores as bi nary objects in DB, to be queried using a new web method (/griddata) –Returns grids (ESRI ASCII format or rendered images) for a given time ran ge or a stream of values for a single cell within a grid (as a time series).  JDBC wrapper –Connects to remote DB directly, without going through standard GSN communication channels –More efficient processing, used in PermaSense  Sensorscope wrapper –The previous Sensorscope wrapper was extended to handle the new packe t format and new sensor types (e.g. Decagon 10HS and Apogee SP-212).  Tweeter wrapper –Enable data fusion between sensor data and Tweeter data in real time

10 Various Technical Support  Data access via web services –A standard web service was developed for accessing data in GSN from exte rnal tools like Geotop and Alpine3D. –A C++ client for data access using user’s existing programs.  Continuation of the access control component in GSN –GSN discretionary access control for each virtual sensor. –Access can be made through web interface (using interactive login) or thro ugh passing of parameters using REST web calls or standard web services. –Direct passing of parameters without going through the interactive web se ssions was needed for automatic tools such as Mathlab or R.  Install, configuration, patches  Continuous development of SwissEx wiki

11 Planned Work  Linking Swissex data –e.g., Linked Open Data (LOD) data –PlanetData (FP7 NoE) consortium –OpenIoT (FP7 strep) –Scalable solution necessary  Completion of ontology-based sensor data search  Cloud-based sensor data management –Mirroring the current data with public/private cloud services

12 Questions ?


Download ppt "Swiss Experiment EPFL-LSIR Report Hoyoung Jeung SwissEx Annual Meeting, Zurich 15 th June."

Similar presentations


Ads by Google