The Department of Human Services has blamed a WAN optimisation tool for an outage to the country’s child support IT system earlier this year.
The outage to both the Cuba and Pluto systems – which run in parallel – caused problems for staff providing child support services or assistance to families for almost a week in March.
The Community and Public Sector Union claimed that the outage left staff locked out of customer files and forced to take messages on pen and paper.
The legacy ‘Cuba’ system is used to provide around $3.5 billion in support payments for 1.2 million children a year whose parents have separated.
It supports the newer staff-facing front end ‘Pluto’ system introduced by the department in July 2017 by providing assessments and calculations.
At a senate estimates hearing late last night acting CIO Charles McHardie said neither system had caused what he characterised as “not a complete outage, but system degradation across the board”.
However the age of the Cuba system, now 16-years-old, appears to have factored heavily in the failure.
“To cut a long story it was not a failure of the Cuba application itself or of Pluto," McHardie said.
Instead he blamed a "WAN optimisation tool" made by Riverbed and "rolled out with Telstra ... that was intended to make our overall wide area network perform faster."
“What these wide area network optimisers tend to do is reduce the amount of latency in the network," McHardie said.
“They also look at transmission control protocol window sizes, so basically make the efficiency of data flowing across the network increase.
“One of the problems that we had with this was that due to the age of the Cuba system – the core products that sit behind the system are known as CA Gen and CICS, which is the customer information control system – had conflicts with the WAN optimisation tool which had this significant impact across the network.”
McHardie said the degradation of service, which began on Thursday March 22 this year, included application crashes, slow application response, and "at time failure for the application to start".
“So the bottom line was we experienced about a 70 percent degradation of child support processing services across the network,” he said.
The department responded with “several days of fault investigation”.
“Once we’d completed our fault finding, which was complete on the evening of Tuesday March 27, we shut down the complete WAN optimisation tools across the network and we restored all of our services back to their normal operating procedures," he said.
The department told estimates that Cuba system would remain in place for some time, but that it was awaiting findings from a Deloitte study before making any decisions on its future.
McHardie said the CICS and CA Gen software that supports Cuba were currently being upgraded.
“... Cuba will be with us for some time yet as we go through the plan to align it with the WPIT [welfare payment infrastructure transformation] program to be able to remove that calc engine, so we need to look after Cuba and ensure it remains fit for purpose for some time yet.”