Government NSW Spatial Services
Every road and place name in NSW, moved onto a new workflow platform without interrupting the service.
NSW Spatial Services replaced the two systems behind the naming of the state's roads and places. We built the new web portal, connected it to Jira as the workflow engine that runs every naming proposal, and brought both name registers across from the previous databases.
At a glance
- Client
- NSW Spatial Services
- Sector
- Government
- Solution
- The NSW Place and Road Naming Proposal System: a web portal on AWS, connected to a Jira workflow engine, with both name registers migrated to PostgreSQL
- Engagement
- January 2019 to early 2020, delivered with OneBlink
- Services
- Migration: about 596,000 rows in all. Key record types included: 23,957 road names, 83,569 place names and 77,132 gazettal records
Context
Spatial Services plays the central role in NSW Government's spatial and land information services. Its data underpins mapping across government, industry and the community. Two of its systems managed the naming of roads and places: a seemingly simple function that matters a great deal when an ambulance is trying to find your address. One handled road naming proposals from councils. The other is the master record of every geographic name in the state, including parks, lakes, mountains, rivers and beaches, and their Aboriginal names and history.
The challenge
Both systems ran on technology that was no longer supported: a road naming system built on an IBM workflow platform and an Informix database, and a place names register on Oracle dating back to 1990. They were unstable, expensive to run and carried a real risk of failure and data loss. Every change to the records had to follow the Geographical Names Act. Users lacked basic workflow, validation and reporting, user management was manual, and the interface did not meet accessibility standards (WCAG).
What we did
A portal connected to a workflow engine
We built the web portal that councils and other naming authorities use to propose road and place names, check them against the naming rules, upload proposals in bulk from Excel, answer queries and follow each proposal through to gazettal. Behind it, Jira replaced the previous IBM workflow platform as the workflow engine. We connected the two: every submission starts or updates a Jira workflow that routes the work between councils, Geographical Names Board staff and the Surveyor-General, and status changes in Jira flow straight back to the portal, so councils always see where a proposal is up to.
Name checks run while the proposal is being entered: offensive words, format rules, and duplicates in the council's own area and its neighbours', using Spatial Services' location data. Councils get a PDF report of the result before they submit.
Two registers, migrated with scripts you can run again and again
The data sat in Oracle and Informix databases in a government data centre, behind a firewall. IBM no longer supported the driver needed to read it, so we wrote our own to extract it. We copied the source databases into PostgreSQL, then transformed them into two new registers, one for road names and one for place names, where every table keeps its own change history automatically. The migration was written to be run and re-run exactly, as many times as needed, and Spatial Services can re-run it themselves.
A modern platform
- A serverless architecture on AWS (Lambda, API Gateway, Cognito, SQS), designed and built in agile sprints.
- A responsive, WCAG-compliant web portal, with automated tests and CI/CD pipelines for simple, safe releases.
The result
- Every road name, place name and gazettal record brought across from the previous databases: about 596,000 rows, with scripts that can be run again at any time.
- No business interruption during the move.
- Still live today: councils propose and track names through the portal at proposals.gnb.nsw.gov.au.
- The code and the migration scripts were handed over to Spatial Services.
Where it led us
This was our first system built around a workflow engine, before we were a Microsoft Dynamics partner. Seeing what a workflow engine could do for a process this tightly governed sent us looking for the most capable workflow platform we could find, and that was Microsoft Dynamics 365. It's also where our migration method comes from: look at the data first, make every run repeatable and account for every record. We bring the same discipline to every Dynamics 365, Business Central and Dataverse migration, now with AI doing much of the profiling and mapping.
We needed a service provider to gain a deep understanding of the challenges we were facing, and to assist us to build a new service that would firstly, overcome those challenges so we could do a lot more than we were previously able, and then provide a platform upon which we could execute some of the key strategic objectives we had for the near and mid-term future. Crucially all of this needed to happen without any business interruption. The team achieved that with aplomb, and we are incredibly appreciative of the role their AWS skills and expertise played in the success of the project.