- Posts: 98
- Thank you received: 0
Internet Speed Test
18 years 6 months ago #14853
by chandak76
Internet Speed Test was created by chandak76
Hi,
Is there a reliable way that I could test the internet sped on my network,I suspect that my isp is sharing my line with oher people,
Is www.internetfrog.com a reliable way to use?
Please recommend any other ways i could use.
Is there a reliable way that I could test the internet sped on my network,I suspect that my isp is sharing my line with oher people,
Is www.internetfrog.com a reliable way to use?
Please recommend any other ways i could use.
18 years 6 months ago #14856
by Arani
Picking pebbles on the shore of the networking ocean
yes leap frog is a good site to test your internet speed.
you said you ISP was sharing your connection. what kind of service do you have, is it internet via cable or adsl? you need to know about this because every ISP shares there resources to give service to their customers.
you said you ISP was sharing your connection. what kind of service do you have, is it internet via cable or adsl? you need to know about this because every ISP shares there resources to give service to their customers.
Picking pebbles on the shore of the networking ocean
18 years 6 months ago #14864
by d_jabsd
Replied by d_jabsd on topic Re: Internet Speed Test
One thing to keep in mind is that no speed test is reliable.
Every test you run will give you a different result. 9 times out of ten, it will report back slower than your actual speed.
Speed tests are great for getting a general idea of what you line is capable of, but there are way too many variables to account for to call a speed test a 'reliable measure of your bandwidth'.
The following variables can skew speed test results-
-Items you can not control:
Routing Changes
Backbone Peering agreements
Speedtest Server hardware
Speedtest software and file size
the number of users currently running speedtests
The speedtest provider's network
-Items you may be able to control:
Your Edge router
Your firewall
Internal network congestion
Speed/Duplex of your devices
Java clients (if speed test is Java based)
Applications currently running.
A better speed test is to convert your data rate to kilo bytes per second. this will give you the max theoretical speed.
Then go about your day, doing what you do, and taking note of the speeds you get when it is displayed.
IIRC, a T1 maxes out at 192 KBps. If you download an ISO image and average 180 KBps, you can be assusred that you are getting the speed you pay for.
I deal with 'speedtesters' on a daily basis. They run speed tests all day long and call for support once a result pops up that they don't like. They never seem to understand that it doesn't mean they aren't capable of using their bandwidth. It only means that they couldn't get mx bandwidth to that particular server at that particular time, due to the large number of variables involved.
As far as sharing resources goes, there is always some sharing at some point in the path. Cable/Wireless is a shared medium all the way to the CPE. DSL/T1/T3/etc. become shared once you hit the provider edge all the way to the internet. This known as oversubscription and is based on the assumption that the customers will not use their max bandwidth all the time at the same time.
Example: A T3 is 45 T1s. A provider will oversubscribe that T3 to provide T1 service to, lets say, 100 customers, knowing that not all of them will peg thier circuit at the same time. Without oversubscription, the ISP would never be able to recoup costs and make a profit.
Oversubscription does have limitations and it requires careful capacity planning to ensure that you can deliver the best service to the customer.
The only time you can reduce the amount of oversubscription is if you buy transit instead of just an internet connection, but then you will have to ensure that you meet the requirements of buying transit (minimum traffic requirements, peering agreements, etc.).
You will still be subject to over-subscription as a transit customer, but not at the same level as an internet customer.
Every test you run will give you a different result. 9 times out of ten, it will report back slower than your actual speed.
Speed tests are great for getting a general idea of what you line is capable of, but there are way too many variables to account for to call a speed test a 'reliable measure of your bandwidth'.
The following variables can skew speed test results-
-Items you can not control:
Routing Changes
Backbone Peering agreements
Speedtest Server hardware
Speedtest software and file size
the number of users currently running speedtests
The speedtest provider's network
-Items you may be able to control:
Your Edge router
Your firewall
Internal network congestion
Speed/Duplex of your devices
Java clients (if speed test is Java based)
Applications currently running.
A better speed test is to convert your data rate to kilo bytes per second. this will give you the max theoretical speed.
Then go about your day, doing what you do, and taking note of the speeds you get when it is displayed.
IIRC, a T1 maxes out at 192 KBps. If you download an ISO image and average 180 KBps, you can be assusred that you are getting the speed you pay for.
I deal with 'speedtesters' on a daily basis. They run speed tests all day long and call for support once a result pops up that they don't like. They never seem to understand that it doesn't mean they aren't capable of using their bandwidth. It only means that they couldn't get mx bandwidth to that particular server at that particular time, due to the large number of variables involved.
As far as sharing resources goes, there is always some sharing at some point in the path. Cable/Wireless is a shared medium all the way to the CPE. DSL/T1/T3/etc. become shared once you hit the provider edge all the way to the internet. This known as oversubscription and is based on the assumption that the customers will not use their max bandwidth all the time at the same time.
Example: A T3 is 45 T1s. A provider will oversubscribe that T3 to provide T1 service to, lets say, 100 customers, knowing that not all of them will peg thier circuit at the same time. Without oversubscription, the ISP would never be able to recoup costs and make a profit.
Oversubscription does have limitations and it requires careful capacity planning to ensure that you can deliver the best service to the customer.
The only time you can reduce the amount of oversubscription is if you buy transit instead of just an internet connection, but then you will have to ensure that you meet the requirements of buying transit (minimum traffic requirements, peering agreements, etc.).
You will still be subject to over-subscription as a transit customer, but not at the same level as an internet customer.
18 years 6 months ago #14866
by chandak76
Replied by chandak76 on topic Re: Internet Speed Test
We've got a 256k line,it sometimes gets too slow and this is usually after hours where theres maybe 2 people in the building,thats why im suspecting that they use some of our bandwidth at those times
18 years 6 months ago #14867
by The_Geek
Replied by The_Geek on topic Re: Internet Speed Test
And also, whenever you use an "off network" testing site, you have to remember that, for example, if you have Charter and run a speed test to a site that is on the Bellsouth network and cross over Verizon to get to that site, if there is any network problems over Bellsouth or Verizon's networks, then your speed results will reflect this, then my phone rings from a customer reporting slow internet.
It's amazing how irate people can get when you prove to them the connection they are paying for is up and running at speed, yet they are slow once they leave our network.
Hindsight being 20/20, I should have stayed in heating and air conditioning.
It's amazing how irate people can get when you prove to them the connection they are paying for is up and running at speed, yet they are slow once they leave our network.
Hindsight being 20/20, I should have stayed in heating and air conditioning.
Time to create page: 0.136 seconds