Title: Interoperability
1Interoperability in-action perspectives from
UK academia
James Reid GeoServices, EDINA
10 February 2005
2Overview
- Who we are
- What we do
- Why Interoperability?
- Interoperability in practice
- Concluding remarks/demo
3EDINA - Who we are
- A National Data Centre for Tertiary Education
since 1995 - based in the Data Library
- Our mission...
- to enhance the productivity of research,
learning and teaching in UK higher and further
education - Focus is service
- e.g. Digimap, EMOL, etc
- but also undertake rD projects ? Services
- e.g. JORUM, SUNCAT, Shibboleth, Go-Geo!
- Until recently, main focus has been provision of
services fund by the Joint Information Systems
Committee (or JISC)
4Research and geo-spatial data team
- Largest team within EDINA
- mixture of GIS specialists and software engineers
- Highly experienced and skilled team
- provides advice nationally and internationally
- active in standards development
- active in GI community nationally and
internationally - First online GI service, UKBORDERS, launched in
1994 - Demands of the services offered means team has
been at leading edge of GI service development in
UK - Strategic move toward interoperability
5What we do - Some statistics
- Digimap
- Until 2002, largest online geospatial database in
the UK (300m objects) - in 1999, it took 70 days to load and convert the
data - 17,000 users (30,000 over 4½ years)
- Average 23,000 files downloaded per month,
200,000 maps generated, 10,000 maps printed off - In 2003, users downloaded over 6.5m worth of
data - UKBORDERS
- 300 boundary data sets
- 70 look up tables
- 1200 downloads per month
- Value to community of key downloads gt 1M
6Corollary of what we do - Service requirements
- Fast servicing of requests
- Scaleable
- accommodates steady or increasing demand
- Robust (our SLD aspires to 98 uptime!)
- Maintainable (see next point)
- Standardized
- Can easily substitute components for repair,
upgrade, etc - Rapid prototyping and rollout
- All above on tight budget ?
- (An aside whats the Business case for
Interoperability Performance? Cost-reduction?
Maintainability? RAD? - recent OGC sponsored research suggests that
saving money is not actually perceived as that
important!!)
7OGC and interoperability
- Open Geospatial Consortium (OGC), a private
sector initiative, formed in 1994 - aim is to develop software specifications to
advance geo-processing interoperability across
the GIS industry - employing practical test-beds and a consensus
specification development process to arrive at
open specifications for standard interfaces and
protocols - defined web service implementation specifications
for - Map Services (WMS) Gazetteer Services
- Feature Services (WFS) Geoparser Services
- Coverage services (WCS) Catalog Services
8The vision - a SDI for the UK academic community
9Data Access - a one-stop shop
Based on R. Wagner 2002
User
WWW-Browser
Go-Geo! Portal
WFS Client
WMS Client
WAAS Client
Clients
Services
WAAS Service
WGS Service
Catalogue Service
WMS Service
WMS Service
WAAS Service
Athens
geoX walk
WFS Service
Geo-Data
Data set 2
Dataset 1
Geo-Data
Security Zone
Research Council Institute
EDINA
JISC Data Centre
10Perceived benefits of Interoperability
- Increases the value of existing and future
investments in Information Systems. - Allows portability of data.
- Expands choices for vendor alternatives no
vendor lock-in. - Enables vertical industry segments to unify
trading practices. - Decreases the long-term cost of ownership for
applicable software investments. - Enables leverage of existing skill-sets, i.e.,
does not require proprietary training. - Provides a benchmark for software design.
11Specific Project aims
- to prove the feasibility of delivering
geo-spatial data using OGC standards - to demonstrate ease of use and value added
- to build support and enthusiasm for further
development - to stimulate and advance further thinking and
- to identify major hurdles in full development.
12Project Outputs
- A range of OGC based web services (WMSWFSWCS)
- A basic annotation web service (XIMA) currently
investigating IBM WBI development kit for Java to
develop a Geoserver (WFS) plugin proxy server
to translate requests - A series of demonstrator clients to illustrate
- Access to data (see later)
- A teaching focussed use case (Metosat data in
teaching weather forecasting) - A research focussed use case (based on dynamic
image registration using web services) - A report on the utility and issues surrounding
implementation of open standards for geospatial
data within the JISC IIE, including an assessment
of security and access authorisation issues
13Data access demonstrator Issues (1)
- Issues
- Identify what OGC web services available
(estimated that worldwide there are only c.250
public WS services and most of these serving
only sample or test datasets) see
www.refractions.net/ogcsurvey - We identified c.20 WMS, 4 WFS, 2 WCS
- Ensure all conform to standards (scale hints
missing, layer names cryptic SRS missing
versioning dialogue issues) - Need for local registry (meta-information)
- How to rationalise users view with disparate
views afforded by different services (may not be
a 11 correspondence of portrayal and data)
ontology? - Layer control and legend issues
14e.g. Legend issues
Example ltLayergt ltNamegtRIVERSlt/Namegt
ltTitlegtRiverslt/Titlegt ltAbstractgtContext
layer Riverslt/Abstractgt ltStylegt
ltNamegtdefaultlt/Namegt ltTitlegtDefaultlt/Title
gt ltLegendURL width"180"
height"50"gt ltFormatgtimage/giflt/Format
gt ltOnlineResource xmlnsxlink"http//
www.w3.org/1999/xlink" xlinkhref"http//glo
be.digitalearth.gov/globe/en/icons/colorbars/RIVER
S.gif"/gt lt/LegendURLgt
lt/Stylegt lt/Layergt
As well as representing legends in different ways
in the capabilities file, the images themselves
can vary in size and style. Problems can also
arise from similarities between legends, where
the same colour is used to mean two or more
things depending on the layer viewed.
15Data access demonstrator Issues (2)
- Issues
- Latency and asynchronicity (especially if doing
lots of round-tripping) - Specification clarity e.g. exact definitions of
some operations in Filter Spec, output schema for
WMS GetFeatureInfo XIMA leaves a lot unspecified
? - Specification harmonisation see next slide.
Addressed under OWS Common? - (04-016r5 e.g. WFS 1.1, Catalog 2.0)
- Metadata and sane names
- Variable quality e.g. granularity and precision
of data - (you pay for what you get?)
16Differences between WFS and WMS capabilities
(Nuke Goldstein Oct 2004)http//www.directionsmag
.com/article.php?article_id686trv1
17Preliminary conclusions
- More work required than possibly initially
anticipated (though overheads with modern tools
is less significant than was required previously
e.g. MMS) - Building the services as well as the clients!!
- Differences in underlying technologies may impact
upon the degree of support for standards (open
source vs commercial) - Leading edge or bleeding edge?
- Security and DRM issues barely addressed how do
OGC web services map into mainstream Web
Serices what about WS-Securitylonger term
where does e-Research and GGF approaches to
security fit in? - Interoperability by definition assumes a minimum
of 2 endpoints providing the services
themselves is only half the story! Still early
days
18Demo
19Interop servers
- ICEDS (http//iceds.ge.ucl.ac.uk/) - A
demonstration service provided by University
College London and ESYS plc, funded by the
British National Space Centre, serving SRTM and
Landsat data at full resolution for Africa, the
Indian sub-continent and Europe. - DEMIS (http//www.demis.nl/home/pages/home.htm)
Company providing range of OGC products and
services - GLOBE (http//www.globe.gov/globe_html.html) - A
worldwide hands-on, primary and secondary
school-based education and science program.
Provides access to datasets for download and a
WMS server. - EDINA (http//edina.ac.uk) National Data Centre
serving UK higher and further education,
delivering inter alia geospatial data and
service, including OGC based ones - IONIC (http//www.ionicsoft.com/) - Company
providing range of OGC products and services.
20Demo Fallback
21Demo Fallback
22Demo Fallback
23Demo Fallback
24Demo Fallback