CIRA’s Internet Performance Test lets anyone with an internet connection test how fast their connection is under real world conditions.

The test lives at performance.cira.ca and takes less than a minute to complete. The end user receives a readout on their download and upload speeds, as well as other quality of service metrics including jitter, latency and packet loss, to name a few. This allows users to compare the performance of their connection against, for example, the service they’re paying for, or that their neighbours receive.

Once complete, the test results are anonymized and added to a national database of internet performance data that provides researchers, policymakers and other interested members of the public with a bird’s eye view of connectivity across Canada. 

Canadians have performed over 1 million tests on the platform since its launch.

Test methods used by different test platforms come in 2 distinct types: “On-Net” and “Off-Net”.

“On-Net” tests generally stay within an ISPs own network and connects to the closest server available – in other words it is a local test. An On-Net test shows the potential capacity of the network - that is it shows your ISP’s network can provide the “up to speeds” advertised by testing your local connection to see if that network is delivering the Internet speeds the ISP sold you.

“Off-Net” tests are what the Canadian Radio-television and Telecommunications Commission (CRTC) refer to as “real world” tests because the test purposely leaves the local network and goes to a Tier 1 Canadian City Internet Exchange Point (IXP) or test server. This is done specifically to show how a person’s overall Internet connection is performing.

When a user initiates a test using IPT, the test leaves their local ISP’s network and travels to the a nearby Internet Exchange Point, and back again. This allows CIRA to measure the actual performance of an Internet connection in real network conditions where users are accessing materials outside their local ISP’s network.

By comparison, tests such as Ookla’s SpeedTest.net generally measure the speeds between the end user and their local ISP – or an “on-net” test. While this is a good way of assessing the top speed you can receive across the ISP’s “last mile” network, it does not necessarily reflect how users access information over the internet under real world conditions.

The CRTC defined measurement methods in Telecom Decision CRTC 2018-241 as, “…these measurements should be performed by an independent 3rd party and measured through a connection from the customer premise modem to an Off-Net server in a Tier 1 Canadian city”.

As defined in this CRTC Decision, IPT meets or exceeds the requirements for this test methodology.

It should also be pointed out that there is no “wrong” test, just different methods, and reasons. IPT is testing the quality of your complete Internet connection whereas most tests focus primarily on speeds inside a localized network. These On-Net tests look at the capabilities of the local network and not the whole connection capability, but again, both test methods are valid although On-Net tests are not accepted by the CRTC.

“Off-net” testing more closely approximates how people use the internet. For example, if a business based in Kelowna, BC wants to video conference with clients in Toronto, ON, their connection will leave their local ISP’s network. Off-net test results will provide the business owner with more useful information about whether their connection is able to meet the organization’s video conferencing and data transfer requirements.

In addition, “off-net” testing has been established as the standard for assessing internet connections by the CRTC (see our answer above or Telecom Decision CRTC 2018-241).

Finally, while popular apps and services like those provided by Google, Facebook and Netflix may be cached within the networks of large urban ISPs, that is much less likely to be the case in rural areas. Rural users will be forced off-net to access popular apps and services. Given the tremendous disparity between urban and rural internet speeds, off-net testing provides rural users more useful information about the capabilities of their connection. 

The CRTC’s project includes more than 2,000 unique testing sites and participation from many of the country’s largest ISPs, but does not include the country’s small, regional providers. In addition, it only includes wireline services (Cable, DSL and Fibre) and does not include fixed wireless services that rural users commonly rely on.

As a result, the experiences of Canada’s rural residents are not fully reflected in their CRTC reporting – experiences that are essential for policymakers to understand where broadband improvements are most urgently needed.

By comparison, CIRA’s Internet Performance Test relies on over 1 million tests from over 270,000 unique testing sites across the country and is completely ISP and/or service agnostic. 

ISED’s National Broadband Map reports data that is provided to them from all Internet Service Providers (ISPs). This data indicates what the ISPs claim is the service available in the standard 25 sq. km Hexagons used by the Government of Canada. This is known as “self-reported data” and is not verified by ISED. It is extrapolated into 250m roadside segments and not by exact locations.

IPT measures actual tests performed by individuals in their homes and offices and is reported right down to the building. It does not rely on information provided by any ISP.

CIRA makes recommendations on how to get the best results.

The most important is to connect your computer directly to the modem/router using an Ethernet cable. Wi-Fi networks are notoriously unreliable for doing Internet testing for a number of reasons, particularly the higher the connection speeds you have and will be challenged to provide the accuracy available from a direct connection to the modem.

We also recommend disconnecting other users and devices, do not run other applications when doing a test, and use a trusted browser like Google Chrome. If you insist on using a Wi-Fi connection, use a modem with a minimum Technical Standard of 802.11a or g for connections up to 20 Mbps, 802.11n for connections up to 100 Mbps, 802.11ac for connections up to 200 Mbps and 802.11ax for connections up to 2 Gbps.

IPT is primarily testing Canada’s Internet to locate unserved and underserved areas across Canada, identify primary areas where better Internet is needed and to track upgraded Internet services to make sure that they deliver what is promised.

IPT test servers are designed to test Internet speeds up to 1Gbps, but really, any speed above 100 Mbps is sufficient for most users and should not require any focus or funding to be improved until all Canadians have access to that level of service. Given that the majority of rural Canada does not have access to 50/10 connections we have a long way to go.

Testing any connection above 1 Gbps may not be accurate for a variety of network environmental reasons and the further you go beyond a 1Gbps connection the less accurate your testing may be. When you try to measure a connection that is 2.5 Gbps or more, the test server will probably fail to run the tests at all. There are built in limiters to prevent very high-capacity tests, so servers are not overwhelmed for users testing much lower connections that need to be fixed. At 500 Mbps speeds or greater, test accuracy should not even be an issue given they are so much faster than most Canadians have access to.

We are in the process of upgrading all IPT test sites to 10Gbps capacity to be able to support even more testing, but this will not change the basic capabilities of the test and its servers as we search for and confirm areas that need improved connectivity.

Although the test servers sometimes will get busy, you should be able to run a test easily in a few minutes. If you keep trying to run a test and it’s always busy, then you probably have a Firewall or Anti-Virus software that is blocking the test. Please try and temporarily disable these and see if it resolves the problem. For example, we have seen quite a few instances of Sophos for MacOS block the test. If you turn off “Web Protection” temporarily then the test will run as designed.

Be sure to turn your Firewall or Anti-Virus back on after you are done running tests.

The CIRA Internet Performance Test will run on any desktop/laptop browser that supports the latest standards. In some situations, accurate speed testing can run at the limits of software and hardware capability and so you may see differences based on non-standard configurations, other plugins and factors beyond the control of the CIRA IPT.

We have tested on Chrome, Firefox and Internet Explorer. We have found in our testing that Google Chrome provides the most accurate and consistent results across all situations, but for most users all browsers will deliver consistent results. 

Here are some quick ways to stop the most common issues that may slow down your connection:

Rerun the test selecting a different test server and try running the test at a different time of the day.  The Internet Performance Test doesn’t test using the optimal path like an ISP test would (on-net); it tests the actual path that your data will typically travel (off-net) as per the CRTC test requirements. Any additional traffic on that path will affect your test.

After discounting the external issues, now you can look at your own home network: 

1. Ensure you're getting an unaffected test. Stop any ongoing downloads and close any programs that may be using your connection.  Running multiple applications uses up your available bandwidth. This can slow down your computer's central processing unit (CPU) and therefore make the speed test inaccurate. To allow the Internet Performance Test to run at its full potential, please close all applications and pause all downloads, video streams, picture uploads, etc.
2. Eliminate external influences. All tests should be run by connecting your computer directly to your modem/router using a standard Ethernet cable for best results and which will also eliminate the possibility of wireless interference. If you do use Wi-Fi to connect, check your Wi-Fi speed and signal strength. Sharing your Wi-Fi connection across multiple devices uses up your bandwidth, which can reduce your overall Internet speed.    
3. Reboot your modem and router and check their capabilities. It's simple, but it might solve a lot of problems. Check your modem’s capabilities. Less capable modems will not handle higher speed connections. Check the label on your modem and look for a Standard designation. 802.11b, for example, will not handle a speed of more than around 5.5 Mbps. 802.11a and g will handle up to 20 Mbps, while 802.11n will handle 100 Mbps.

After performing all these steps, you should have enough evidence to contact your ISP if there's still a mismatch between the speeds you’re expecting and the speeds you're getting. See if they have additional suggestions or if there are diagnostics, they can run on their end.

The speed at which you send files from your computer to an Internet site. For example, the speed it takes to upload photos.

Mbps or “Megabits per second” is a measurement of speed on the Internet. It is equal to one million bits of data transferred per second.

This map provides a quick view of your results in comparison to your neighbours, other cities, provinces and across Canada.  By selecting Download, Upload, IPv6 or DNSSEC off the dropdown list, you can view the various results of the Internet Performance Test.  Zooming in or out on the map, dragging the map or doing a search on a desired postal code will quickly let you view the details you wish to see.  Hovering the mouse over a particular hexbin will provide the details for the selected results for that area.

IPT performs a multi-step process to set your location and postal code:

1. Check IP Address in a GeoIP database to determine your approximate location. Use that rough location to set the Postal Code.
2. Ask the browser for its geo-location attributes. This must be approved by the User. If they do approve the use, then we use the geo-location from the browser to set the Postal Code and the latitude and longitude.
3. If the Postal Code (in left hand side of test) is wrong the user can change it to a correct Postal Code. This will get the user closer to their actual location.
4. When a User clicks on “Test” they will be able to refine the location to its best granularity by dragging the marker pin that is on the map to their exact location.
5. Once set, the Postal Code and Location will be remembered in a Cookie.

So, each step gets the location more and more accurate.

Very rarely when running a test and moving the location pin on the map to your home or business you might receive this following message when you attempt to run the test:

If you entered your correct Postal Code then simply select the first option, “Use the postal code I entered: xxx xxx” and the test will complete using that Postal Code. If you find that the Postal Code offered in the second option is the correct one, then select that option and the test will complete. If you find that you have incorrectly positioned the test marker then select the third option, reposition the pin and complete the test.

Note that if you had positioned the test location pin correctly and selected the first or second option, you should get the same message each time you run a test and will have to select the option again.

To run the performance test, we look up your IP address in a GeoIP database to determine your ISP.  In most cases this is an accurate query, but if it is an incorrect result, the test allows you to enter the correct ISP. 

The CRTC established the Universal Service Objectives or minimum services each Canadian should have:

• Download Speed at least 50 Mbps.
• Upload Speed: at least 10 Mbps.
• Latency/ping: a round-trip latency threshold of 50 milliseconds (ms).
• Jitter: a threshold of 5 ms or less.
• Packet loss: a threshold of 0.25% or less.

DNSSEC applies digital signatures to incoming DNS data to scan for authenticity and to verify its integrity. Broad adoption of DNSSEC can protect domain names from attacks by hackers. Adoption of DNSSEC by organizations, applications and Internet Service Providers will make Canadian Internet more secure in future. If the test is indicating a green checkmark for DNSSEC, your ISP supports this security extension. A red 'X' indicates your ISP is not yet supporting this feature. Read more about DNSSEC here.

IPv6 is the newest version of the Internet Protocol, the language that computers use to communicate with each other over the Internet. To secure Canada's Internet future, Canadian organizations need to get on board with the next generation Internet protocol, IPv6, which will make available millions of more IP addresses. If the test is indicating a green checkmark for IPv6, your ISP is ready for the future! A red 'X' indicates your ISP is not IPv6 enabled. Read more about IPv6 here.

The Internet Performance Test analyzes more than 100 variables to help determine if there are issues within the network and your connection to the internet.  These variables and their results are displayed once a test is run and the Advanced button is clicked.  To read more information on those advanced results, click here.

To run the performance test, we look up your IP address in a GeoIP database to determine your province and based on that we use the closest server. 

Downloading from a local server is generally faster than a distant one, so selecting a different server prior to running the Internet Performance Test may result in slower speeds.

CIRA highly recommends that you always connect your computer directly to the modem/router using an ethernet cable. Wi-Fi networks are notoriously unreliable for doing internet testing for a number of reasons. Particularly, higher connection speeds will be challenged to provide the accuracy available from a direct connection to the modem.

Wireless routers, depending on quality and type, can have low signal strength. This might cause you to experience slower speeds and following are the technical standard numbers found on your router and their theoretical and actual for speed (Mbps) capabilities. Only rely on the actual capacity noted in the following table.

The advanced details provided in our test can be used to pinpoint specific issues that may be resulting in poor performance. There are several sites and public forums that discuss speed tests as well as how to improve performance using these data as talking points.  A good start is DSL Reports.

The data you create from each test will be anonymously aggregated with all the other performance tests done throughout Canada. The data is sent back and forth to an M-Lab server.  Together, these results help researchers understand and build a better internet for Canada.  This data does include your IP address along with the results of your test.  All the data is made publicly available in order that researchers can access this information to perform analysis.  For an example of the type of output created from these results, see the visualizations section of the M-Lab site.

http://www.measurementlab.net/visualizations

The IPT does not collect personal information on M-Lab servers, such as your Internet traffic, your emails or Web searches.

M-Lab provides the largest collection of open Internet performance data on the planet. As a consortium of research, industry and public interest partners, M-Lab is dedicated to providing an ecosystem for the open, verifiable measurement of global network performance. Real science requires verifiable processes and M-Lab welcomes scientific collaboration and scrutiny. This is why all of the data collected by M-Lab's global measurement platform is made openly available and all of the measurement tools hosted by M-Lab are open source.

While the data from this specific test is not available for use yet, M-Lab does provide access to their data which will include data from this test as well as other similar tests that use the M-Lab infrastructure.

Future phases of this Internet Performance Test will allow access to the data that has been gathered through this specific test.

To access the data currently available thru M-Lab, check out the following links:

Data collected by NDT is available at: https://www.measurementlab.net/data/

Internet performance tests take into account a variety of measures and employ different methods of testing. Often, results from other speed tests will provide varying results for the same users. Read our blog post, How does the CIRA Internet Performance Test compare to Speedtest for more information.

Are you curious to know what is planned for upcoming phases of the Internet Performance Test?  Visit the development roadmap for the future of the Internet Performance Test.

The advanced details provided in our test can be used to pinpoint specific issues that may be resulting in poor performance. There are several sites and public forums that discuss speed tests as well as how to improve performance using these data as talking points.  A good start is DSL Reports.