Go Big or Go Home: How Much GIS Infrastructure Do We Really Need?

Presented by:
David Nicosia - Remme Corporation

Transcript

1. Go Big or Go Home: How Much GIS Infrastructure Do We Really Need? Dave Nicosia Remme Corporation
4. Custom Project Tracking Software Project Origination Component
5. Map image courtesy Bing Maps www.bing.com/maps Contact Information Job Location “Project”
6. “I can handle the user interface, I’m just not as familiar with storage options for the web.”
7. High Level Requirements •  Collect Project Text Attributes •  Collect Project Geographic Features •  Store Project •  Recall Project •  Render Project as KML •  Email notice of new Project request Browser Browser Browser / Server Browser / Server, ??? / Server Server Server
8. Project Text Attributes – Jobsite address – Contact name, email address Geographic Features (0 – m) Point features for comments (1 – m) Line features representing horizontal bore holes – (1) Point feature for origin – (1) Point feature for destination Start End
9. Tool Bag Approach Browser – OpenLayers (2.x, 3) – Leaflet – Google Maps API – esri Javascript API – MapQuest Javascript API – Bing Maps API Server – TinyOWS + PostGIS + Postgresql – Featureserver – Flask + spatialite – PHP – CGI or FCGI + Python, Perl, BASH, C, Fortran, …
10. Integrated Solution Approach Browser + Server – ArcGIS for Server – Google Maps Engine – GeoServer – OpenGeo Suite – GeoDjango
11. Hosted Solution Approach Browser + Server – ArcGIS Online – Google Maps Engine – CartoDB – Giscloud
12. Other Considerations •  Existing infrastructure •  Technologies we already know •  Number of expected end-users and anticipated load •  Ongoing cost (licensing, service fees, etc.) •  Security •  Reliability
13. High Level Requirements •  Collect Project Text Attributes •  Collect Project Geographic Features •  Store Project •  Recall Project •  Render Project as KML •  Email notice of new Project request Browser Browser Browser / Server Browser / Server, ??? / Server Server Server
14. High Level Requirements •  Collect Project Text Attributes •  Collect Project Geographic Features •  Store Project •  Recall Project •  Render Project as KML •  Email notice of new Project request Browser Browser Browser / Server Browser / Server, ??? / Server Server Server
15. Store Project Recall Project To KML Email Notice TinyOWS + PostgreSQL + PostGIS Yes Yes Yes No Featureserver Yes Yes No / (maybe via OGR) No ArcGIS for Server Yes Yes Yes Maybe ArcGIS Online Yes Yes Yes Maybe Google Maps Engine Yes Yes Yes Maybe GeoServer Yes Yes Yes No OpenGeo Suite Yes Yes Yes No CartoDB Yes Yes Not really No PHP Yes Yes Yes Yes CGI (Python) Yes Yes Yes Yes
16. “I need a geospatial database”
17. Data volume Performance Concurrent access Spatial indexing Geometry operations
18. Web Browser Web Server Geo DB or or
19. How much GIS infrastructure do we really need?
21. Further Considerations PostGIS –  system PostgreSQL –  how to add PostGIS spatial extensions (permissions, version) –  system PostgreSQL updates PHP –  new to me –  additional libraries (spatial, JSON, KML, email) System Python environment –  add libraries like Psycopg2 or MySQLdb1 –  system Python updates
22. High Level Requirements •  Store Project •  Recall Project •  Render Project as KML •  Email notice of new Project request
23. Web Browser Web Server JSON HTTP CGI Python Programs
24. “Project” JSON Representation { “id”: “xxxxxxxxxxxxxxxxxxx”, “name”:”Bob Smith”, ”email_address”: “bob@bobsconstr.net”, “job_address”: “200 Congress, Austin”, “comments”: <GeoJSON FeatureCollection>, “conduits”: <GeoJSON FeatureCollection> }
25. High Level Requirements •  Store Project •  Recall Project •  Render Project as KML •  Email notice of new Project request cdsdb.py cdstokml.py cdsemail.py
26. HTTP Resources •  Store a new Project POST projects/ •  Overwrite a Project POST projects/<id> •  Recall a Project GET projects/<id> •  Delete a Project DELETE projects/<id> •  Recall a Project as KML GET askml/<id> •  Email a new Project notice GET notice/<id>
27. cdsdb.py import os import sys import json import shelve
28. cdsdb.py operations 1 client – 4.5 txn/sec 2 clients – 9.1 txn/sec 8 clients – 36.5 txn/sec
29. cdstokml.py import os from cdsdb import Storage from String import Template
30. cdsemail.py import os from cdsdb import Storage from String import Template import smtplib
31. Map image courtesy Bing Maps www.bing.com/maps Contact Information Job Location “Project” Submit Job Request Notice (via Email) Project as KML Save/Recall Project
32. Reminded to… •  Appreciate and use existing infrastructure •  Think critically about requirements •  Seek simplicity
33. Thank you. Questions or Comments?
34. “You should have used a database.”
35. “You should have used PHP.”
36. “You should have…”