Download - Why Change ? An overview of ‘OpenGIS’ management issues Derek E Reeve University of Huddersfield
Structure
• Going Mainstream
• What is wrong with what we’ve got?
• The ‘Enterprise Computing’ Vision
• Alternative Open Systems strategies– Universal Storage
– Universal Access
– Universal Delivery
• Goodbye GIS?
Going Mainstream
‘You could go so far as to say that the GIS industry as it has historically been defined is going to be
gobbled up by the mainstream IT infrastructure…
… I get the sense at the moment that Microsoft have bigger fish to fry than spatial data but it is only a
matter of time’
Stafford A, CEO Smallworld, 1997.
Going Mainstream
‘GIS used to be a boutique industry, but now it’s going mainstream. Anyone who doesn’t keep up
with this development won’t last long’
Kim K. (president of Harvard Mapping and Design), 1997
Going Mainstream
• Influences– Technological convergence
– Commercial convergence
– Organisational convergence
Going Mainstream• Technological Convergence
– Dual architecture GIS is no longer justified
– New database technologies
– Standardising influence of Windows
Going Mainstream
• Technological Convergence
– Standardising influence of WWW
– ‘Shrink-wrapped’ GIS.
– ‘COTS’ GIS
Going Mainstream
• Commercial Convergence– The GIS vendors are minnows compared to the
major Information Systems companies, such as Oracle and Microsoft.
– The major companies will dictate the future of IS technology.GIS will adopt technology standards set by the major companies
Going mainstream
• Commercial convergence– GIS companies need alliances with major IS
companies to survive
‘Over the last five years we’ve come to the realisation that we can’t do everything. We need to plug into existing information systems… A company that fails to integrate its system close to the leading edge is a company that can easily be left behind’
Boyle, v.p. Intergraph 1997 quoted in Wilson J
Going Mainstream
• Organisational Convergence
User organisations expect to integrate geo-spatial data into mainstream workflows.
What is wrong?
• Novel data types
– Relational model designed for alpha-numerics– New data types - graphics, video, sound, email,
html, etc not accommodated.
‘80% of corporate data are NOT held in DBMS systems’
Baker B, 1997.
What is wrong?
80 % of data is NOT in DBMS!
80% not in database
•complex
•unstructured
20% in database
•simple
•structured
What is wrong?
• Islands of automation– information flow discontinuities
– interruption of workflows
– specialist databases
– proprietary systems
What is wrong?
• Islands of automationDomain Specific
DataTabular
Data
DatabaseDatabaseApps.Apps.
GISGISApps.Apps.
GIS MIS
DatabaseServer
Source : Steiner J, Oracle Corp 1997
-Specialist data held in separate specialist systems
- Data isolation
- Islands of automation
- disruption of workflows
What is wrong?
• Over-specification– ‘Bloated’ applications
– Over sophistication of functionality
– Excessive costs
– Restricted deployment
What is wrong
• Management complexity– complex support functions
– specialist support skills
– upgrade/maintenance functions
– High hidden total costs of ownership (TCO)
What is wrong?
• Excessive development times– business is getting faster
– traditional information systems development cycles are too slow.
Enterprise computing
‘a digital nervous system for an organisation’
‘Information at your fingertips’
‘Raising the organisational IQ’
‘Everyone’s Information System’
Enterprise computing
• The Vision– Data accessible across and between
organisations– No platform boundaries– No data format boundaries– No vendor boundaries– More efficient, easier administration etc– GIS fits into this or will be marginalised
Enterprise Computing
• The ‘Universal’ technologies :-
– ‘Universal’ Storage
– ‘Universal’ Access
– ‘Universal’ Delivery
Enterprise computing
• Universal storage– ‘Universal Servers’
• Oracle Universal Server • Informix Universal Server IUS • IBM DB2 UBD • Sybase Adaptive Server enterprise
– Extended RDBMS, not OODBMS.
– ‘Cartridges, Blades, Extenders’
Enterprise Computing
• Universal Storage
• Benefits – spatial data held as part of core corporate data– RDBMS facilities :- scalability, back-up,
integrity, security, roll back, recovery etc– integrated support and administration etc
Enterprise computing
• Universal Servers
MISApplication
Records
Integrated Spatial Information Management
Desktop GIS
Client
WebBrowser
Proprietary SpatialData
Oracle7
Spatial
Serving Information to Users
ImageText
Source : Steiner J, Oracle Corp 1997
Enterprise Computing
• Universal Storage
• Problems– Legacy systems will not die easily
• investment in ‘legacy’ systems
• staff skills
• established work practices
Enterprise computing
• Universal access
– An alternative strategy• Leave the data where they are
• Construct a ‘virtual database’
Enterprise computing
• Universal Access : ‘Interoperability’
– 3 levels :-• Hardware : ‘Plug n Play’
• Data : Location independence
• Software : Componentware assemblages
Enterprise computing
• Universal access – Key features
• Three-tier architectures
• Middleware
• DCPs
Enterprise computing
• Universal Access : Three tier model
CAD GIS RDBMS Text
‘Virtual’ Relational DatabaseMiddle ware
Components
DCE
Open systems
OGIS
etc
User User User UserUser
‘L e g a c y’ d a t a b a s e s y s t e m s
Enterprise computing
• Universal access– Key features
• ‘Wrappers’
• Components
‘Components put the engineering into software engineering’
Enterprise computing
• Universal Access : GIS– OpenGIS Consortium
– New Generation GIS products• e.g. Geomedia
Enterprise Computing
• Universal Access : GIS– GIS Componentware
• e.g. Geomedia
• Map Tools
• ‘Put a Map in your App’
Enterprise computing
• Universal Delivery– Problems with client-server databases
– Benefits of WWW and Browsers
– Intranet and Extranets
Enterprise Computing
• Universal Delivery : Web GIS
– Geo-spatial data needs to be integrated into the Web environment
– Alternative strategies• thick browsers
• thin browsers
Enterprise computing
• Bringing it all together– the three ‘Universal’ strategies are compatible
• Universal servers can be part of a universal access network
• Web browsers are simply a universal front-end
– together they will deliver most aspects of the ‘enterprise computing’ vision.
Conclusion
• Goodbye GIS?– The future is interoperable
– Geo-spatial data processing is integrating into this future
– the end of GIS or ‘GIS in disguise?
ReadingBaker W., (1997) Universal Access vs Universal Storage. Thirteenth International Conference on Data Engineering, IEEE,Birmingham, April 7-11.
Buckley D. J. (1998) The GIS Primer : An Introduction to GIS (http://blaze.innovativegis.com/education/primer/primer.html)
Buckley R. (1998) Two Products with one Name, Mapping Awareness, July 1998, 29-33.
Corbley K and Buckley R (1998) The Best of Both Worlds Mapping Awareness, June 98, p22-24.
ESRI (1997)Enterprise Spatial Data Management with SDE. ESRI ARC News, Summer, 1997, 8.
Maguire D J (1995) A High performance, Object-oriented Spatial Database Management System. AGI'95 Conference Proceedings, AGI : London, p 1.22.1-1.22.4
Mapping Awareness (1997), Who's Who in GIS Number 4 : Smallworld Systems, Mapping Awareness, 11,10,p15.
O'Leary A. (1997) Managing Mission-Critical Text Byte September, 1997, 43-44.
Smith R., (1997), How to move GIS into the mainstream. Proceedings AGI'97 conference, Birmingham.
Stonebraker M. with Moore D., (1996) Object-relational DBMSs : The next great wave. Morgan Kaufmann, San Francisco.
Theriault D and Buckley R (1998) Share the data wealth Mapping Awareness, March 1998, p18-21
Wilson J. D. (1996) Enterprisewide Implementations Transcend GIS Boundaries, GIS World, 9,10, 42-48.
Wilson J. D., (1997), Technology Partnerships Spark the Industry, GIS World, 10,4,36-42.