The Covid-19 crisis has resurrected some ancient unresolved issues in the tech industry, like peering which everyone thought was long resolved.
While the industry should fix the problems lack of peering causes, the acrimony around the issue is entirely unwarranted.
Peering refers to peers or providers connecting their networksdirectly to each other, often at a convenient central exchange point. It can be done settlement free, or for an agreed charge to cover costs of building out network capacity.
For end users, peering is an esoteric issue with a name that is misleading as providers connecting to each other might not be peers, but very different in size and network reach.
There are good technical reasons for connecting directly. In simple terms, peering can help boost performance and network responsiveness as it shortens and increases the paths that internet traffic takes.
Peering can also reduce providers' costs as they can connect to others at points of presence instead of buying separate circuits to each telco (for example) so as to reach their customers.
Chorus, which was demerged from what used to be Telecom, is keen and sees peering as a business opportunity.
The network infrastructure company is working with the New Zealand Internet Exchange (NZIX), which is the country's largest peering provider, owned by its members.
"Our belief is that getting traffic closer to end customers on the network means a better experience for those customers.
A better experience means they will likely take up new internet services with added traffic and grow the utilisation of our network," a Chorus spokesperson told me.
Describing the peering concept as a win-win for the industry, Chorus hopes it will make providers use more of its EdgeCentre colocation facilities as this will mean they won't need a separate site to peer which in turn will reduce their costs.
Covid-19 crisis permitting, the Chorus - NZIX trial will result in a commercial service in the last half of the year. The trial is a natural extension for Chorus, an infrastructure wholesaler which makes money out of connecting internet providers.
Nevertheless, from the above it should be apparent that there is also a strong business case for not peering if you're a big player like a telco incumbent.
NeedToKnow3
Last week there was an ugly public spat between small internet providers and Spark which, like Chorus, was born out of the Telecom split.
Spark does not and has never peered with the rest of the .nz internet. Until December this year, Spark and Vodafone were the two non-peering holdouts. Now it's only Spark, which charges other providers for connecting to them.
This is said to be causing issues for broadband customers connecting to others on Spark's network, as the increased usage with video conferencing and working from home during lockdown has eaten into available capacity.
At first glance, that seems really callous but in reality the situation is more complicated.
First, Spark has no obligation to start peering and to reduce other providers' costs at the detriment of its own business. Even so, Spark has boosted domestic network capacity for its wholesale customers by 40 per cent since March, for free.
Providers that don't pay to connect directly to Spark often do so in Sydney instead.
Connecting to each other over 2000km away from New Zealand seems strange, but Spark peers freely in Sydney. While it's suboptimal to connect that far away instead of in Auckland, it doesn't add much delay and network capacity between Australia and New Zealand costs far less nowadays than it used to.
Across the Tasman the shoe's on the other foot for Spark. In Australia, Spark is a minnow compared to TPG, Telstra and Optus which do not peer with anyone. Spark has to pay to connect to the big Aussie telcos and it's been like that for decades (and angered smaller providers there too).
The important thing here is that there's zero chance of Spark changing its corporate mind and starting to peer with other providers in NZ at this time.
It is unreasonable in the extreme to demand that Spark take a massive risk and reconfigure its network to start peering in the midst of an extremely serious and worsening global crisis.
The routers that forward user data packets could crash as complex configurations change, and smaller providers' equipment might not be able to handle a large amount of new destinations for traffic.
Such mishaps are not unheard of. They would create an uproar in New Zealand because Spark handles so many customers' essential network traffic. The XT mobile network fiasco is still fresh in Spark's memory.
That said, it's important for everyone that the internet works well during the pandemic. Accusing Spark of profiteering from Covid-19 is just nasty and unhelpful however. How about everyone steps back, talks to each other calmly and collaborates to fix the problems instead?
