NBN Co corrects national address database

 

Maker defends against inaccuracy claims.

The National Broadband Network rollout will improve the centralised address database that has so far forced NBN Co to double check for inaccuracies, according to the Government's key spatial information agency.

The G-NAF database, used by government agencies and private companies alike, provides addresses for 13.14 million Australians premises collated by PSMA Australia and based on data collected from state and territory governments, Australia Post, the Australian Electoral Commission and other agencies.

Up to 50,000 new or modified addresses are added to the database every three months but the quarterly update schedule has proved difficult for users of the database, which include emergency services.

NBN Co chief executive Mike Quigley told a Senate estimates hearing last month that up to 30 percent of data for apartment blocks, provided under a two-month contract with PSMA, had proven inaccurate.

Though the accuracy of data for single homes was much higher, at an estimated 95 percent, Quigley said the inaccuracies had exceeded NBN Co's risk assessment and forced it to double check the estimates.

"[PSMA data] isn't quite what we need to do this job," he said at the time.

The Office of Spatial Policy, which oversees much of the geospatial data used by government agencies, acknowledged problems with the existing database.

""There is some element that is causing the NBN a little bit of grief. We understand that," said Helen Owens, OSP general manager.

"Certainly there is no better geocode national address file than the G-NAF, currently.

"[G-NAF] doesn't contain addresses to a level where there are gated communities or there is a building with multiple stories, or there are multiple dwellings in one location. How it all comes together is very complicated."

However, she said the double checking undertaken by NBN Co on the address database would feed back to PSMA, ultimately improving the database for all government users.

Quigley told a radiocommunications conference this week that the company had worked with Queensland-based industrial mathematics firm Biarri – signed in July last year – to re-check address data from PSMA and correct any inaccuracies.

The data was being used primarily to discover the ultimate cut-off points used between the three technologies planned for use under the NBN.

NBN Co is serving 93 percent of mostly urban premises with fibre, four percent of regional areas with wireless and the most rural four percent for satellite.

However, Quigley said it was "exceedingly difficult to figure out where those cross-over points are".

"The problem is you're out at the tails of the distribution and it's a very difficult model; if anyone thought it was an easy job to plot it out and know where the cross-overs are, that's a mistake," he said.

"One of the worst cases I had was the G-NAF database told us there were 150 G-NAFs - that's people where we've got to put a premises down - and when we actually had a look at it, it was in the middle of a paddock."

According to Owens, NBN Co's corrected data would ultimately be used as an additional address data source for PSMA, which is planning to launch a web portal that would give real-time access to new or modified addresses as they are collected and cleaned.

She said the agency had sought to sign a whole-of-government agreement with PSMA for use of the database.

"Ultimately the NBN will give us a geocode address file that is very very accurate," Owens said.

Copyright © iTnews.com.au . All rights reserved.


NBN Co corrects national address database
 
 
 
Top Stories
Five zero-cost ways to improve MySQL performance
How to easily boost MySQL throughput by up to 5x.
 
The big winners from Defence’s back-office IT refresh
Updated: The full list of subcontractors.
 
Tracking the year of CIO churn
[Blog post] Who shone through in 12 months of disruption?
 
 
Sign up to receive iTnews email bulletins
   FOLLOW US...
Latest Comments
Polls
Which is the most prevalent cyber attack method your organisation faces?




   |   View results
Phishing and social engineering
  69%
 
Advanced persistent threats
  3%
 
Unpatched or unsupported software vulnerabilities
  11%
 
Denial of service attacks
  6%
 
Insider threats
  12%
TOTAL VOTES: 985

Vote