Building reliable redundancy for Westminster
Andrew Moore-Crispin • September 13, 2017if( has_post_thumbnail( $post_id ) ): ?>
We don’t think much about how the Internet gets to our homes.
Well, I mean, we here at Ting do. We think about it all the time.
We, in the collective sense, as citizens and users of the Internet, don’t. Unless our access to the Internet goes away. Then we have some questions that deserve answers.
Since Ting brought the first fiber customer online in Westminster back in August 2015, we have faced some challenges. Some of those challenges have become network outages. We’d say more outages than we’d like but one outage is more than we’d like.
Network growing pains.
What have we done to fix the issue?
As of today, construction, installation and lighting of the new redundant backhaul connection for the Westminster Fiber Network is finished. A redundant backhaul connection is basically an alternate route to the larger Internet. It’s what will prevent unavoidable issues, like another provider or contractor accidentally severing a fiber line (by far the biggest issue we’ve had to contend with in this context) or a squirrel that needs more fiber in its diet from chewing through a line (which happened yesterday) from becoming an extended outage.
What is a redundant backhaul connection?
A redundant backhaul connection is basically a detour that Internet traffic can use if there’s a problem. A sideroad that accepts the overflow in the case the main highway is closed because of an accident, to follow the metaphor.
The Internet is, as the name suggests, an interconnected network. It’s distributed and there’s no single point of failure for the network at large. There shouldn’t be a single point of failure for a network node, like the fiber network here in Westminster, to reach the larger Internet… and now there isn’t.
We’re officially redundant in Westminster. That’s great news.
Can we promise 0 outages forever? No provider can do that and to be frank, during the “dirt and hard work” active construction phases, the risks of something breaking are obviously greater. What we can promise is that we’ve built the appropriate infrastructure that will step in to keep local problems from becoming network-wide issues, and that we aren’t satisfied with good enough either.