home bbs files messages ]

Forums before death by AOL, social media and spammers... "We can't have nice things"

   alt.os.linux      Getting to be as bloated as Windows!      107,822 messages   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]

   Message 107,516 of 107,822   
   Lew Pitcher to All   
   Re: How do "they" Speed-test Internet Li   
   08 Sep 25 13:43:27   
   
   XPost: alt.comp.os.windows-11   
   From: lew.pitcher@digitalfreehold.ca   
      
   On Mon, 08 Sep 2025 23:24:12 +1000, Daniel70 wrote:   
      
   > On 8/09/2025 10:21 pm, Dan Purgert wrote:   
   >> On 2025-09-08, Daniel70 wrote:   
   >>> [...]   
   >>> Which got me thinking ..... How do "they" Speed-test Internet Links?? ..   
   >>> particularly How do 'they' distinguish the Up-link TIME from the   
   >>> Down-link TIME?? e.g. my current speeds, using Speedtest [...]   
   >>   
   >> For the most part, a speedtest works by you downloading a file of known   
   >> size (say 100 MiB), and then sending it back.  Exact file size will vary   
   >> by testing provider, but essentially it's just this:   
   >>   
   >> Download start = 0.00   
   >> Download end = [TIME]   
   >>   
   >> File size / TIME = X Mbit / sec   
   >   
   > But how does the distant end know when I have received the entire file   
   > (i.e. Download end time)??   
   >>   
   >> Upload Start = 0.00   
   >> Upload end = [TIME]   
   >>   
   >> File size / TIME = Y Mbit / Sec   
   >>   
   > Similarly, how does the distant end know when my computer started the   
   > Upload (Upload start time)??   
      
   The transfers typically use TCP, and that protocol includes feedback   
   from the receiving end as to whether or not it has received the data   
   sent. Typically, this feedback is in small enough packets that transmission   
   latency doesn't affect the overall throughput measurement enough to   
   matter.   
      
   In theory, only one end has to make the timing measurements; It can   
   be the sending end, or the receiving end.   
      
   For sending-end measurement, the sender   
   - starts the clock   
   - sends a known amount of data to the receiver, obeying the receiver's   
     TCP flow-control requests   
   - stops the clock at the TCP FIN/FIN-ACK end-of-data acknowledged   
   - computes elapsed time   
   - computes upload transfer rate (known amount of data / computed elapsed time)   
      
   For receiving-end measurement, the receiver   
   - waits for the TCP connection   
   - starts the clock   
   - receives the data, counting the volume as it goes   
   - stops the clock at the TCP FIN/FIN-ACK end-of-data received signal   
   - computes elapsed time   
   - computes download transfer rate (counted amount of data / computed elapsed   
   time)   
      
   With those web-page throughput estimators (https://www.speedtest.net, for   
   example),   
   the throughput computation usually occurs at the web server, with the webpage   
   providing   
   javascript (or other code that will execute on the client system) to either   
   sink the TCP connection and data being sent (from the web server to the client   
   system   
   for the upload test, or source the TCP connection and data (from the client   
   system to   
   the web server) being sent for the download test.   
      
      
   HTH   
   --   
   Lew Pitcher   
   "In Skills We Trust"   
   Not LLM output - I'm just like this.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]


(c) 1994,  bbs@darkrealms.ca