The verdict is in. Telstra’s July network crash was the result of a series of oversights and poor architecture choices over the last six years. But while the review by an external consultancy tells us everything about how the fault in the emergency call system developed, we still really don’t know why.
It’s not clear that the CEO and the board know either. CEO Vicki Brady said in a statement the outage “was primarily a result of us not treating network timing as a critical capability within the network (or a ‘sovereign function’) requiring the highest levels of oversight and protection.”
So someone decided that network timing, foundational to any IP network, was not a “critical capability?”
The consultants, Technology Audit Partners (TAP), said modifications made in 2020 “degraded the mobile core timing architecture,” introducing the potential risk of “looping” – that is, an infinite loop of network layers confirming and re-confirming the wrong time. Tellingly, TAP added, there was “no evidence” anyone involved with this change was “aware of or investigated this risk.”
It’s not just Telstra that has had trouble keeping mission-critical infrastructure going.
Outages aren’t rare
Rival Optus was last year fined 12 million Australian dollars (US$8.7 million) and now faces a AU$250 million ($180 million) court case over an emergency call system outage that resulted in four deaths.
Globally, network crashes among big telcos aren’t rare either, a paper by six Australian academics reminds us. It points to T-Mobile and DT in 2020, KDDI and Rogers Communication in 2022, AT&T in 2024, Telefónica in 2025, to name a few.
The paper, published in IEEE Access a year ago, argued for a shift in IP network management from prioritizing redundancy to mitigating complexity. “By examining the vulnerabilities inherent in critical IP functions and services, such as DNS and BGP, we have underscored the potential for minor misconfigurations to precipitate large-scale network outages,” the authors said.
But while the sheer demands of giant IP networks may be a factor for all telcos, the other side of this is what Telstra has been prioritizing. Since 2018 the company has been driven by successive multi-year plans intended to cut costs, simplify its products and improve customer service.
The strategy has certainly boosted the bottom line. In 2018-19, it reported earnings of AU$2.15 billion (US$1.55 billion) against revenue of AU$27.8 billion ($20 billion).
Seven years later it has maintained net income at AU$2.4 billion ($1.7 billion) even though revenue has shrunk 18% to AU$22.9 billion ($16.5 billion). The tight financial management is reflected in the 24% gain in Telstra’s stock price over five years.
But it’s impossible not to connect the dots between the cost-cutting and the weak network oversight revealed in the TAP report. Whether it is a result of deprioritization, a lack of engineering expertise or simply over-stretched staff, we don’t know.
But it should be a cautionary tale for all telcos trying to staunch the narrowing cost-revenue delta while relying on IP networks that are growing in complexity and centrality to the entire telco business.
