3. 1. Introduction
· This work consists in the deployment (and
un-deployment) of SQL scripts in
transactional WPS servers at runtime.
· For the concrete implementation, we used
the geoprocessing SQL scripts provided by
GearScape and the 52º North
implementation of the transactional
extension of the WPS services (WPS-T).
5. 2. GearScape geoprocessing language
· SQL'92 standard implementation together
with the OGC simple feature access
specification
· Independent from format, file type and
location of source data.
· Local execution
7. 2. Advantages
· SQL language is simpler than imperative
languages.
· SQL scripts can be encapsulated as black
boxes and reused later by using wizards.
· SQL language can be seen as a language to
'speak' about geoprocesses.
9. 3.1. WPS
· Set of standard operations for the
publication and execution of any geoprocess
via web.
· Defines 3 operations:
- GetCapabilities
- DescribeProcess
- ExecuteProcess
10. 3.2. WPS-Transactional
· WPS standard does not define how to
deploy (and undeploy) geoprocesses at
runtime.
· 52º North Initiative proposes an extension
with 2 specific operations for that purpose:
· DeployProcess
· UndeployProcess
· 52º North Initiative itself has developed an
implementation of the extension.
12. 4.1. Description
· WPS-T implementation has been provided
the ability of 'understanding' GearScape
geoprocessing language.
· As a result, the server can describe and
execute existing SQL scripts as well as
deploy new SQL scripts at runtime.
13. 4.2. Workflow
· Create an SQL script with (or without)
GearScape editor.
· Deploy the script in a running instance of
the 52N WPS-T server implementation.
· Reuse the geoprocess by executing it using
any WPS client.
17. 4.3. Sample scenario
- As proof of concept 2 scenarios have been
created:
· Reusability
· Interoperability
- The screencasts can be found in the
Documentation section in
www.fergonco.es/gearscape
18. Future work
- Raster data integration in the SQL engine.
- Sextante algorithms integration in the SQL
engine.
- Parallel computing in the SQL engine.