On Friday July 8th around 4:45AM, the Rogers network in Canada went down hard.  Cellular, internet, and cable services were disrupted for customers across the country.   The NetBlocks monitoring group estimated the outage knocked out a quarter of Canada's observable internet connectivity.  And this wasn't a short outage - it took Rogers almost the entire day Friday to restore core services, and until late Saturday night for some customers to be fully restored, including myself.  

The impact was felt across the country - calls to emergency services failed, Interac payments were down, concerts were cancelled, and countless businesses had to stop work for the day.  And this isn't the first time in recent memory Rogers has suffered a major outage - In April 2021, Rogers experienced a country-wide outage that saw many unable to place calls, send text messages, or access the internet via mobile devices.  

Since the outage, there have been calls for "more competition" to mitigate the impact of a single carrier failing, but the answer isn't that simple.  We have competition today, through the CRTCs Wholesale access regime, but that regime also failed on Friday as third party ISPs such as TekSavvy were also impacted when Rogers went down.  This is because while companies like TekSavvy lease the "last mile" from Rogers, the current configuration of how they interconnect depends on the Rogers network being functional for that wholesale access to work.  

One solution to the problem is to break the incumbent carriers up through a process called structural separation.  As an example in this model, Rogers would be broken up into several companies - one that operates the physical network (cable, fibre, etc), one that operates the wireless network (LTE), and one that provides services to end-consumers.    This is the model employed in the UK, where BT Retail provides broadband services to residential and business customers, and Openreach provides access to network infrastructure used by competitors to deliver their own broadband products.

In this model, an outage of the Rogers "retail" network would not have impacted the wholesale services provided by the likes of Teksavvy, since an outage on the retail side would have no impact on the operations of the infrastructure provider, providing for greater resiliency to single points of failure.  It also fosters competition as the physical network operator is encouraged to partner with wholesalers to help fill the network to capacity.

This isn't a radical idea.  In Singapore, they deployed a national broadband network using an open access network model.  The physical fibre is operated by a company called NetLink.  This entity designs, builds, owns and operates the passive fibre network infrastructure.  This entity then provides open, wholesale access its fibre network, so that telecommunication operators can focus on offering innovative products and services to consumers and businesses without incurring high fixed costs of building fibre networks.  It's win-win - with a single wholesale provider powering the backbone of the fibre network, Singapore avoids the need to overbuild fibre to the premises, while still allowing for competition through the open wholesale access model.

As I've said many times before, the problem with the CRTCs focus on facilities based competition leads to inefficient network builds.  Take any new residential build in Canada -  at least two providers (CableCo and TelCo) will both build physical fibre to the same house to provide services.  In an ideal world, we would have one provider lay the physical fibre, and then any number of providers accessing that network on a wholesale basis to provide services to end consumers. Some might say this provides for lack of diversity, but you don't have two gas lines coming into your home, nor do you have two electricity feeds - you have one and then a choice of which service provider delivers the end-service to your home.  

So maybe it is time we take a serious look at structurally separating the incumbent networks and separating them into infrastructure companies and service companies - the results could be amazing for true service based competition in Canada.