Hello all
An Open Geospatial Consortium (OGC) meetings happened last week in
Orléans. Below is a few notes on topics that attracted my attention (of
course much more were discussed during the meeting):
Which OGC standards for the cloud?
One topic discussed during the meeting was execution of user-provided
algorithms on the cloud, for avoiding the cost of transferring large
amount of data to the user machine. We had an update of NOAA Big Data
Project [1] with three questions:
1. What OGC standards are still relevant when data are hosted (and
computed on) in the Cloud?
2. Which OGC standards are not relevant when the end user is also in
the Cloud?
3. What new/revised OGC standards are needed?
An other initiative facing the same questions is openEO [2], also
presented during the meeting. That initiative proposes an API in Python
which can be executed on the clouds. It seems to have similar goals than
GeoAPI, with a different approach:
* GeoAPI follows closely OGC/ISO abstract models, while openEO seems
to define their own model.
* (Partially a consequence of above) GeoAPI defines low-level objects
(metadata, referencing) and is progressing toward high-level objects
(features, coverages) while openEO starts from high-level objects
(image, /etc/).
Those questions can be discussed on a GitHub issue [3]. Note that the
Google Earth Engine given in example seems to use a mechanism similar to
Java Remote Method Invocation (RMI); even if the syntax is different, we
can recognize stubs, /etc./
The GeoAPI meeting presented a first draft of metadata API in Python
[4], in complement to the already existing Java interfaces. We plan to
provide two proof-of-concepts, one using GDAL-Python bindings and one
using Java-Python bindings (the later will enable the use of Apache SIS
in Python for instance). Another email specifically on GeoAPI will be
posted later.
Blockchain
Another session discussed Distributed Ledger (Blockchain) technology.
The "OGC essentials" standards are used for recording locations. The
aims to to provide cryptographically signed and verified
presence-claims. Some usage examples are birth certificates, land or
house ownership, and access to a property. The presenter expressed a
need for an international standard, in part because smart contracts
needs to be able to reason about location. Some work are progressing in
a new International Organization for Standardization group, ISO/TC 307
[5] (as a reminder, the group for geospatial information is ISO/TC 211),
but ISO/TC 307 does not seem to include location information yet. Some
company proposes to work with OGC [6], but I'm not yet aware of a
communication channel where the discussion happen.
Others
In addition to above, ad-hoc meetings were held to discuss the use of
Non-Authoritative Data, Earth Observation Exploitation Platforms,
Interoperable Simulation and Gaming, best practices for JSON, and
Statistics. They were also talks about "fog computing", as a variant of
"cloud computing" where the computers are local devices instead than
servers. An example is autonomous cars collaborating together for
resolving an urgent problem, where there is no time for communicating
with distant servers.
Martin
[1] http://www.noaa.gov/big-data-project
[2] http://openeo.org/openeo/news/2018/03/17/poc.html
[3] https://github.com/ogcscotts/TC-Meeting-topics/issues/28
[4]
https://github.com/opengeospatial/geoapi/tree/master/geoapi/src/main/python/ogc/metadata
[5] https://www.iso.org/committee/6266604.html
[6]
https://blog.foam.space/crypto-spatial-coordinate-standard-and-the-open-geospatial-consortium-technical-committee-dffb0bd14c7c
