Understanding Download Speeds
The Internet is a very crowded place. As a result, it is not surprising that performance-related service problems are common and users are concerned whether their Internet Service Providers are delivering the speeds they paid for. Many speed-testing websites are available, either hosted by Internet Service Providers or by independent third parties. Unfortunately, however, not all speed tests are equal and understanding how download speed tests are conducted will prove beneficial to any Internet user.
Other People Are Reading
-
Vehicle Traffic Analogy
-
Internet download speeds can be understood by comparing them to road speed tests. Road speed tests can be taken based on either the local authority assessment of the passenger throughput of a road at maximum capacity, the number of passengers in a car or the speed limit of a road. Each of these measurements is valid in the context in which the test is conducted. Likewise, Internet download speed tests relate to the application and capacity around which the tests are taken. These can be either application speed tests or capacity speed tests.
Application Speed Test
-
This is the actual end-to-end speed that a TCP connection is able to achieve during a real user experience. Since a wide range of applications use connections differently, a download test will not usually invoke a test that matches the application usage, and thus it will fail to deliver the actual performance of a connection relating to that particular application. Thus, a video may take long to download even when the download speeds allow for faster download speeds.
-
Capacity Speed Test
-
This test can be compared to having multiple buses full of passengers on a road. This test simply tests the user's Internet highway, testing the volume of data that can be transported and that arrive at the destination. This test does not take into account the data order of packets sent and thus does not reflect the real speed that an application will be run at. This is because applications will send data in ordered packets, thus limiting the bandwidth usage of any connection.
Route Variance
-
Since speed is measured in time and not in miles, random speed testers on the Internet will deliver varying speeds because of their geographical locations. Reliable speed testers, however, will use different applications in different worldwide locations and compute the difference in results.
Timing Error
-
Since timing in a time-sharing multitasking PC environment is difficult to compute, especially for fast connection speeds, this can sometimes result in big differences. A difference in 100th of a second can result in test reports of 50Mbps for a connection running at 55Mpbs.
-
References
Resources
- Photo Credit web button image by creative from Fotolia.com
