The recent failure of the 111 emergency service — no, not for the new XT 3G mobile network, but for Telecom’s landlines in Auckland and Northland — drew attention to the fact that the incumbent operates the Public Switched Telephone Network (PSTN) on equipment that was installed in the early 1980s.
In the 111 case, the failed piece of gear was the POP4 NEC NEAX 61E switch (or digital phone exchange if you like) that suffered what Telecom called “a software issue”. This in turn saw emergency calls failing to connect for around four hours that day. It’s not the first time POP4 failed: the switch fell over in August 2008, too, again taking 111 emergency calls with it.
What’s more, the gear will be in the PSTN, also known as Plain Old Telephony System (POTS), for another decade at least, making it up to 40 years old before it is decommissioned and replaced with newer stuff.
First, it has to be said that Telecom isn’t alone in keeping gear from the eighties in its network. Most other big telcos around the world are only reluctantly ripping out the digital switches they installed in the seventies and eighties at huge expense.
At one stage, it looked as if New Zealand would be if not the first then an early PSTN replacer, with Telecom having a starting date of 2012 for its Next Generation Network that would see voice customers on a brand-spanking new IP network.
That’s now been pushed back at least eight years. In 2009, Paul Reynolds said in Telecom’s annual report that the company expects to migrate all New Zealand customers from today’s legacy PSTN to an all Internet Protocol network by end of 2020, in line with its legally binding undertakings.
Customer migration will start this year, however, so we’re talking about a decade of shifting to the new platform.
It’s easy to see why: in the face of constantly sliding calling revenues, pushing through a business case to replace the PSTN is an uphill task. Figures such as $300 million capital expenditure have been bandied about, and you don’t get that kind of money from your board unless you show there are revenue streams to cover it. So far, none have shown up.
Meanwhile, Telecom’s Technology and Shared Services team that was led by chief transformation officer Frank Mount, who resigned recently, has to support the legacy technology in the network.
We’re hearing stories about the NEAXs being at the end of their official vendor support because of their age, and how Telecom is having to scrounge for spare parts by buying decommissioned units from overseas telcos. Keeping the magnetic tape drives that feed the NEAXes their software going seem to be a delicate and difficult task as well.
It’s certainly not easy to find much information about the NEAXs or even people who’ve worked on them, but we’ve learned a few things from some retired engineers. For Telecom and Alcatel-Lucent to source the staff with the knowledge and experience to maintain systems like the NEAXs must be similar to banks having to pay top dollar for retired COBOL programmers.
Telecom is arguably in a bind here, having signed a contract with the government that forces it to provide basic phone and dial-up service to people, as well as emergency calling for the country.
Plus, Telecom has promised the government a shiny new voice network while it maintains at great cost a creaking old one that’s prone to failure.
Ten years from now we’ll be able to tell how Telecom fared in this Herculean transition task. It’s hard not to have some sympathy for the incumbent’s staff trying to piece it all together.