Today I was pinging random websites, and I decided to ping a website, from which I was downloading at the speed of 2mb/s, and the ping(latency) was 10 ms.

My question is, if the answer from the server comes for around 10ms, how are these 2mb/s calculated? Any hints, links to explanations or similar are highly appreciated :)

  • Did any answer help you? if so, you should accept the answer so that the question doesn't keep popping up forever, looking for an answer. Alternatively, you could provide and accept your own answer. – Ron Maupin Aug 7 '17 at 18:02

Your ping time latency is how long it takes to get from one end to the other, and it has nothing to do with bandwidth which is how many bits are transferred per second.

Think of latency as the distance between two cities. It takes a fixed amount of time to drive between the two cities. Bandwidth is like how many people you can fit in each vehicle traveling between the cities; a bus has more bandwidth than a two-seater car. It takes the bus and the car the same amount of time (latency) to get from one city to the other, but the bus delivers more people (bandwidth) in that same amount of time.

| improve this answer | |
  • 1
    I like the explanation with the car, but what is interesting to me is how this is turned into network packets. I think that 1gb file cannot be sent in one frame, I am not sure if this is the proper word to say it, but I know that large files are cut into small separate pieces and sent through the network, so I would like to know more about those small pieces and how they calculate the total upload/download speed between two devices. – Kristiyan Katsarov Jan 16 '16 at 20:37
  • Bandwidth is merely how many bits per second are delivered. It doesn't matter how many pieces you group the bits into. A discussion of frames, packets, and segments is really far to large for this site. You need to learn about things like the OSI model and the layers first. – Ron Maupin Jan 16 '16 at 20:42
  • Yes, I know about the OSI model and was wondering at which layer is the bandwidth limitation done, as well what is the relation between the bandwidth and the ping, because if the ping is the time two buses go from one distance to another, what if one bus needs to travel 100km and the other 500km? – Kristiyan Katsarov Jan 16 '16 at 20:52
  • As I wrote, the ping latency has no relation to bandwidth. The primary bandwidth limitation is layer-1, but it gets a lot more complicated beyond that. The buses deliver the same amount of people (bandwidth), but the latency of one is five times as long as the other. Latency and bandwidth are two separate things. – Ron Maupin Jan 16 '16 at 20:59
  • Let's say you want to download 1GB file from one server. The ping to this server is 10ms. Another server (same hardware, different location for example) has the same file, but the ping is 2ms. Let's say there are no bandwidth limitations from ISP, nor the servers have any other traffic, except you. Logically, the download from the second server should be faster as you get the response much faster from it. I think I just melt my brain :) – Kristiyan Katsarov Jan 16 '16 at 21:06

Everything that has been written is true, but if you're using TCP, there is a strong correlation between latency and bandwidth.

TCP does not wait for an acknowledgement to send the next packet. It's using a window. When the window is full, the peer must wait for an ack. Moreover, the faster ACKs are received, the faster the window can grow (it's the slow start algorithm).

The stability of the latency can also come into play.

| improve this answer | |
  • Latency doesn't have anything to do with bandwidth. Just because I may have a large latency and I use TCP doesn't change the fact that my link has a bandwidth of 100 Mbps. My bandwidth doesn't change. It may take longer to download a file with TCP and a higher latency, but the bandwidth remains constant; I can transfer 100 Mbps. – Ron Maupin Jan 20 '16 at 18:58
  • Absolutely correct. The link bandwidth is fixed. However, it seems to me that the question is more oriented towards practical bandwidth, as the number of bits that have been transferred divided by the time it took to download a file on a website. – Xavier Nicollet Jan 20 '16 at 19:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.