You need broadband, and you want some built in redundancy. You’ve got a decent service level agreement (SLA), which your provider has determined based on statistical analysis and advice from actuaries. If you go with two providers you’ll get built-in redundancy: if one triple-nine provider goes down you’ll be fine, because the other SLA also has three nines, and what are the chances of both going down at the same time?
It should be almost impossible, with 99.9 per cent uptime promised during the 0.1 per cent that you’re down. It’s always possible, however, that two different providers are backhauling over the same network, meaning that you’ll get hit with that 0.1 per cent at exactly the same time. Suddenly it’s not a default for uptime, but a redundant failure, which makes the entire effort pointless.
Dave Dobbin, president and chief executive officer of Toronto Hydro Telecom, argues that for systems to be truly redundant they have to be completely independent.
“We call our network a ‘Bell-free zone.’ We can do two entrances, two POPs (e.g. points of presence), two networks. We don’t use Bell ducts.”
Mark Tauschek, senior analyst with Info-Tech Research Group, generally agrees with Dobbin’s argument. “If Allstream buys the copper from Bell, it all terminates at the same location.”
What to do? Tauschek recommends that if you have a circuit from Bell, and another from Allstream, then you should request termination in a different location, or on a different line card or chassis.
“It’s important to get as much diversity and redundancy as possible, because Bell owns all the copper.”
Dobbin can argue that Hydro doesn’t outsource any functions, and that all of its technicians are security cleared with criminal record checks, but the problem with the Toronto Hydro approach remains its cost.
“It’s more expensive initially to lay stuff in,” says Dobbin. “But our business model is predicated on owning your own network. If the customer is large enough to want to buy this, then odds are we can help them.”
Size is the issue. Although Dobbin claims that Toronto Hydro Telecom targets customers with as few as 25 employees, the costs make this primarily a big business technology.
“To bring this to a building costs a fortune,” says Tauschek. “This is a huge amount of money to have to dig up the roads and lay the fibre. Either you have to spend a fair chunk of money yourself, or you need multiple tenants.”
One option is to go with a company like TeraGo Networks, which will put a base station on the roof of your building and take that last mile to the air. The cables are then dropped down – certainly a lot less invasive, and cheaper, than ripping up a street.
TeraGo Networks’ president and chief executive, Bryan Boyd, is also quick to promote the benefit of an independent system.
“What makes us unique is that our infrastructure is completely separate from Bell, Telus, and others. We never touch down on their grid. The best way to mitigate risk is to deal with multiple carriers, and this is where we come in.”
After connecting a building to the TeraGo network there is a ring (or sometimes a number of rings), that circle a city. Between cities TeraGo rents fibre from multiple carriers. Says Boyd, “With our technology there is very little speculative capital expenditure, and ROI can be calculated more easily.”
There are, however, limits to how much diversity you can hope for. If you’re putting in an MPLS infrastructure you’ll need to go with one provider, because there isn’t carrier interoperability. Toronto Hydro Telecom’s network seems a good bet for this: it is provisioned on MPLS and can offer a 50-millisecond failover.