Telstra this morning defended the performance of its mobile network following outages in February and March.
The telco said that, despite different primary causes for each of its recent outages, the large scale disruptions were the result of multiple devices trying to simultaneously re-establish connection to the network, overloading key components of the end-to-end system and exacerbating the impact of the initial outage.
Telstra’s head of networks, Mike Wright, told a press briefing: “The ability of mobile networks to deal with these mass re-registration events is not unique to us. It is a global issue and something our experts are telling us is being faced by many in the industry.”
Wright said that while each of the outages had had different triggers, “the common characteristic is the ability of a lot of the devices to reregister on the network.”
“That is how cellular works and has worked from day one,” he said.
“Devices log on for voice and log on for data… some of the mechanisms and the way they interwork with each other did not manage the traffic into end.”
For example, the March outage was precipitated by a failure in an international cable link used to support overseas visitors roaming onto the Telstra network.
This fault caused large numbers of overseas mobiles to try and re-register on the network, and the resulting overload impacted local users whose devices also tried to re-register compounding the problem.
Telstra’s chief operations officer, Kate McKenzie, said the basic traffic handling capacity of the network was not the problem and this had been amply demonstrated on its free data day on 3 April when a record 2686 terabytes of data were downloaded, a 46 per cent increase on the volume during Telstra’s previous free data day in February.
One user downloaded a massive 994 gigabytes. “That’s as much as the average user would download in 40 years,” McKenzie said.
Wright said Telstra had made a number of changes to fine-tune the system and he believed these combined would reduce the likelihood of a similar occurrence in the future: “There are multiple elements in the traffic management chain.We have made a number of adjustments and we believe the system is far more capable of handling the reregistration mechanism.”
Increasingly, Wright said, Telstra is using data analytics tools to identify issues with its network and to optimise performance.
“Social media is now one of our inputs. Technology can tell you things, but sometimes so too can your customers,” the telco executive said.
Telstra will also increase the processing power of the home location register – the master database that contains details of every subscriber and which is used to authenticate devices and allow them access to the network.
However Wright said this was a precautionary measure, not an indication that limitations in the register had contributed to outages.
He refuted criticisms that Telstra has cut back on internal resources leaving it less able to deal with major network issues: “We have all the capabilities we believe we need and I continue to be very proud of our engineers, but we’re certainly not saying that we cannot get an external review.”
McKenzie said that Telstra had initiated a review of its network using its vendors, and independent experts.
“We are very keen to bring a fresh set of eyes from someone not connected to the network, so we are working with Ericsson, Cisco and Juniper as well as our own internal engineering experts to give an end-to-end review and while that is underway we are managing any changes in the network carefully,” she said.
“We expect to review to be completed in a month or so, not much longer.”
She added: “We have an independent expert, Dave Williams, an international expert with expertise in this type of network. He’s associated with Tech Mahindra and he is providing us with a resource to do a triple check, to see is there anything we have missed. Are we looking at the right things? Is there some stone left unturned?”
Williams will report to McKenzie.