The first phase of the .CA Internet Performance Test officially launched Tuesday, May 5th. Within a week, over 1500 individuals launched the test, giving us a preliminary idea of the interest, effectiveness and the future roadmap for our product. It was a quiet beta period before the storm of public release.
The purpose of this test was to create something uniquely Canadian. We want Canadians to better understand how their Internet connections work – instead of a vague cloud, all Internet users in Canada can learn how their data travels and where it gets stuck. To understand that their real speed is not the maximum speed that their ISP offers them. This is why our Internet Performance tests speed, but also evaluates the overall quality and performance of Canadians' Internet on Canadian Internet Infrastructure. And we can see where things are getting stuck, causing latency or otherwise not working as they should. Armed with this information we can also help encourage more Canadian organizations, governments, and ISPs to locally peer and improve Canadian transit – because all of the Internet matters.
The first weeks were eye-opening. Two people at CIRA, including me, ran the test from on our home connections and realized immediately that we weren't getting the upload speed we were paying for. My colleague called her Internet Service Provider and got a technician onsite to fix the problem. For me, it was an opportunity to upgrade my legacy service to get optimal performance from my ISP for less money! So it actually worked to help a couple Canadians identify problems and fix them. A great start!
With every great start comes a hiccup. M-Lab's open source Network Diagnostic Test (NDT) is deployed globally and forms the basis of our IPT. Everything M-Lab does is open source and for the betterment of the Internet. Linking Canada and Canadians into this fit naturally with CIRA's mandate, which includes doing good things for the Internet.
But open source software has its challenges. One of the biggest hurdles, in this early stage of the IPT, is the reliance of M-Lab's NDT on Flash. Following the beta run of the test -- circulated to our employees, people in their immediate networks and those who found us through online search – we quickly came to realize that Google Chrome is the browser most optimized for Flash. Results coming from tests run using Firefox, Safari or Internet Explorer still worked great, but for those with very, very fast connections or lots of bloat-ware in their browsers, the ability for the browser to execute Flash became the bottle neck and not the Internet.
But we want our Internet Performance Test to offer value to all Canadians – including those who prefer using browser other than Chrome and also those on Apple and mobile. And make no mistake, people are passionate about this subject. We're so happy to receive feedback from Canadians who take the test, or those who want to take the test but don't have Chrome and Flash to get the analysis.
The other challenge with Flash is that we understand Canadians are prolific smartphone and tablet users, opting to use mobile devices over laptops and desktops for Internet-related activities at home. So a broadband test (note not intended to be a mobile test) that works best on a PC is less than ideal. We want to better understand which Canadians are getting better quality and speed of connections on their home and work networks, where and why there is discrepancy. Expanding our platform to be more inclusive of mobile technologies in future phases of the IPT will help us to garner more comprehensive data.
The realization has us excited about the next phase of the NDT, which will use web sockets, allowing us to extend browser support via HTML5 and to broaden the usage of the tool to Apple and mobile users. Most of the above was actually known. Releasing something like a test on a subject about which so many people are passionate and that relies so heavily on open source and crowd sourced information is guaranteed to generate positive and negative responses. One of our biggest surprises came courtesy of a few tech support emails from the good citizens of Calgary.
For our test we are using a crowd-sourced postal code database (yes such a thing actually does exist) and when those postal codes were entered in “reverse” on Calgary's grid-like streets it caused the latitude and longitude of the open source map software to reverse the streets. If you haven't been to Calgary, there are lots of intersections with names like “26Ave SW and 26 St SW”. While Calgarians think it is endearing, we think everyone would be happier if they just named their streets like everyone else does. Until that happens we will work with the folks in the other open source communities to get it cleaned up – and you can too! It is what makes the Internet awesome.
Last updated on December 1 2016.