ESA > Join & Share > WCS Scenario Workshop DLR

WCS Scenario Workshop DLR

HMA Online Data Access Workshop 9 June 2010, DLR Oberpfaffenhofen

"Building no. 122 (new EO Center with glass hall), room 0136 in the ground floor"

Invited: HMA AWG and HMA-FO Task 3 Consortium

Proposed Draft Agenda

09.30 Welcome (DLR)
09.45 Workshop Objectives, final agenda (all)
10.00 Online Data Access WCS 2.0 overview - Peter Baumann
11.00 EO primary data access as planned in SAFER (GMES Emergency Response Service) - Torsten Heinen
11.30 Spot Image Online Data Access requirements - Didier Giacobbo
12.00 Questions and issues for backend connection (ESA)
12.30 lunch
14.00 EO primary data access by the German Center for Satellite Based Crisis Information (ZKI) operated by DLR - Tobias Schneiderhan
14.30 Lessons learnt in implementation and integration of EO primary data access services at DLR - Torsten Heinen/Stephan Kiemle
15.00 Discussion
16.00 Overview of RIDs and Requirements (all)
Demo of the http://www.jeobrowser.com/ client towards the disaster charter catalogue and INSPIRE catalogue
16.30 Conclusions
17.00 Teleconference with Bob Banik - CSA
17.30 Adjourn

Meeting notes (taken by Y. Coene)

HMA-FO Online Data Access Workshop
List of participants
- Stephan Kiemle (DLR)
- PG Marchetti (ESA)
- Peter Baumann (Jacobs University Bremen)
- Gerhard Triebnig (EOX - HMA FO Task 3)
- Christian Schiller (EOX - HMA FO Task 3)
- Stephan Meissl (EOX - HMA FO Task 3)
- Tobias Scheiderhan (DLR)
- Jean-Pierre Gleyzes (CNES)
- Jerome Gasperi (CNES)
- Michael Schick (EUMETSAT)
- Torsten Heinen (DLR)
- Bob Banik (CSA) via Teleconference at h 17.00 CEST
- Yves Coene (Spacebel - HMA FO Project Management Support)
- Andrea Biancalana (ESA - Engineering Support - GIGAS Support Action)

Note actual agenda re -arranged as follows:

9h45: Introduction Stephan Kiemle and PG Marchetti
Objective is to understand relationship between specs for ODA and G/S infrastructure.

CNES: what is the direction we have to follow to not get lost in the many standardizations underway related to online data accesss. If WCS is the direction to go the HMA AWG has to issue a recommendation - loudly!

9h50: Presentation Peter Baumann

slide 8: UML AbstractCoverage with a domain, rangeType and rangeSet

ESA: how does this relate to our actual product formats ?
EUMETSAT: have you looked at ordering specification as for subsetting options, there also the SWECommon is used.
DLR: how does this data model relate to the other OGC data models ?
ESA: how can Capabilities return identifiers of millions of “products” ? A: we will propose a grouping structure, possibly similar to EO WMS solution.
DLR: you need a bridging between coverages to scenes and products. Link resolvable in which direction ? DLR: Depends on server chaining , e.g. once found a coverage, you may want to order the products it is composed of.

Slide 15: subsetting is now “trim” (keeps #of dimensions) or “slice” (reduces #dimensions).

CNES: how return pixels in the GML, you assume WCS returns everything in GML ? A: there are 4 ways in GML for the pixel payload, one can be a separate file with the pixels. Pure GML, pure exchange format, or mixed with a GML wrapper. CNES: Pleiades will use JPEG2000 which contains the metadata instead of the metadata including the pixels (opposite approach).
EUMETSAT: for netCDF, it is sufficient to have the data format to derive the metadata. A: S. Nativi will provide the extension for netCDF.
Slide 17: link with other OGC specifications
DLR: how do we need to understand all these extensions ? Can we choose for EO ?
DLR: modular approach is nice, but how about the interlinking ? For defining “nil” values, they are different in netCDF than geotiff. A: the details are still to be worked out when the work is being done. DLR: do you put the dependencies in the “application profile” ? Where do you state the dependencies between the profiles ?
DLR: can you explain how a datamodel extension for the “core” could look like for nil values ? A: not yet. We will look at sweCommon.
ESA: if we use “jpeg2000”, should we understand that both GML extension and nil extension values are not required as they are included in the format. A: we do need it anyway to help the client to understand that e.g. 255 is nil.
DLR: becomes complicated with many choices: nil value in GML and jpeg2000 may then not be consistent.
ESA/DLR: may be a problem if both the WCS and the products both allow to specify nil values ad the information is redundant and may diverge.
EUMETSAT: nil value should not be mandatory if the format does not allow it. A: the list of nil values can be empty, thus can remain mandatory.
DLR: I thing we need an extension to the core data model for each extension.. A: We are not prepared to discuss this yet (not yet prime-time). One solution may be to group all images under a single coverage (with one dimension less) – to have only one identifier in the capabilities instead of millions - and use subsetting to get the individual images (using time dimension). ESA expects this problem to be analysed and a solution proposed for the MTR review.

ESA: there are a number of open points from this discussion:
- unique identifiers
- question mark on having mandatory or not the nil values in the EO AP.
- Scenario identified by Jerome, client who has identified a product ID and how to transfer this to the WCS?

11h30: DLR Online Data Access Lessons Learned – Torsten Heinen

- Slide 7: DLR use cases for online data access
- Difficult to slide vertically through GEOTIFF files. Alternative storage structure may speed up this.
- DLR: our customers don’t want to retrieve full files and do processing, they would like basic processing and subsetting to be done by the (WCS) server.
- ESA: can WCS be used to return overlays of two coverages ? A: no, but WCPS could be used to do this.

Use of clouds, private clouds, virtual archives is discussed. DLR considers virtualization mandatory.

12h30: Presentation Didier
- ESA: do you see need for a seamless interaction between archive and WCS. A: more important is to have a qualified archive (i.e. specific for the customer/market). Direct access to archive maybe useful for scientific users, but not for commercial users.

14h00: Presentation Tobias Schneiderhan “Emergency Mapping and Disaster Monitoring”

- The Feasibility Analysis, Tasking and Data access time takes today on average 3.5 days. ODA should contribute to decrease this time lost.
- In emergency context, product will be fetched from WCS within 8 hours (rule of thumb).
- Subsetting possibility is not really required: “nice to have”
- EOX: SAFER result output is PDF files. Is there a need to provide output also as layers ? A: We have heterogeneous users, some very basic, other very sophisticated GIS experts.

14h50: SAFER presentation
- demo of SAFER SDI (using Geonetwork)

15h30: CNES presentation of WMS/CSW client (in Javascript).
- integration with Streetview, Openstreet, gazetteer, geoRSS, Flickr etc.
- WCS access and time series (WMS) is planned.
- Currently used for disaster charter client
- Uses proxy (web server) to avoid cross-scripting issues.
- http://www.jeobrowser.com

16h00 demo earthlook.org by P. Baumann

16h10: ESA presentation HMA-FO ODA Use Cases and Scenarios
- ESA: issue: we have a set of documents from Task 3. In WCS documents, requirements for ATS are very difficult to link to the high level requirements. How to solve this issue to have requirements to link to the specifications ? Is particularly important when designing the EO Profile of WCS. (2) should there not be a higher set of requirements from WCS Core to link to the project requirements? Without requirements, we take the risk of a random walk towards an EO WCS profile.

Action: Task 3 to propose a solution for MTR.

ESA conclusion on scenarios:
- (1) most agencies do no see a direct link between archive and WCS. (i.e. systematic production scenario). In this case, WCS is kind of replacement of rolling archive.
- (2) Spotimage scenario: separate server for a community
- Enhancement of the “View services” is required: i.e. “community visualization”.
- (3) Other scenario of a “custom coverage” on a separate (WCS) server.
- Eumetsat would like to link ODA with the “ordering service”, want to give users freedom to subset what they ordered, including for bulk order.
- There is also the end-to-end scenario (from CNES) where access via a simple product identifier to WCS is required.
- Eumetsat, Spotimage: we need to be able to identify the user accessing our data via ODA (is a requirement on “reporting”). Spotimage also wants to limit access to some datasets to certain users.
- ESA: for sentinel – only required to know who has accessed the data. DLR: there is also high level terrasar-X data for GMES, for which authorization is important as well.
- ESA: We have the assumption that security can be decoupled from the service, but may depend on the binding.
- ESA: it is unavoidable that the requirements work is continued/improved.
- DLR: there is no extension foreseen for REST ? Is a modern way to access resources on the Web. A: a volunteer is required to write the extension. Eumetsat: REST may not be easy to integrate with the security.
- DLR: plugin in a new datamodel or the extension of the datamodel ? There should be a metamodel for EO products: products, scenes etc.
- ESA: are all the things part in the WCS Core required or should some things become optional ? e.g. GML wrapper. Is a question for the agencies.
- DLR: traceability between the requirements and the solution you propose is to be improved.
- ESA: workshop of today has identified the important and lower priority scenarios. Furthermore we have seen that the feasibility analisys and tasking contribute to the final data access elapsed time. The GSTP DREAM project will address the feasibility and tasking based on HMA FO outcome.

Teleconference with Bob Banik CSA
- ESA: summary of the workshop presentations and discussion
- CSA: question concerning Sentinel 1 use of HMA. It is agreed that a debriefing will be organised at next GSCB which will take place on 16 June 2010. HMA is the baseline for implementation of interoperability.

HMA FO Mid Term Review is confirmed 5-6-7 July in ESA-ESRIN, Frascati starting mid day on 5 July, & will see an open and closed session of the HMA AWG on 7 July 2010, terminating at h. 16 CEST.
HMA FO Deliverables page

Contributors to this page:

Page last modified on Monday 14 of June 2010 09:27:38 CEST by .