View Single Post
  #25  
Old March 24th 17, 11:23 AM posted to alt.comp.os.windows-10,alt.comp.hardware.pc-homebuilt,alt.windows7.general,alt.comp.os.windows-8
Paul[_32_]
external usenet poster
 
Posts: 11,873
Default Win 10 giga-update: KB 4013429 vs KB 4015438

Andy wrote:
I asked what you got for speed as a question char not what i get


A metric that is easy to understand, is to post a Speedtest.net result.

For example, mine is 15/1, and that's exactly what I pay for.

It wasn't always like that. My first ISP sold 5mbit ADSL, and
delivered 3mbit, mumbling excuses about SNR (which weren't
actually true) as the reason for their clumsy "traffic shaping".

*******

When you download from a server, the server may not max the link
with a single connection. If I download a Linux LiveCD from uWaterloo,
it might only use 1/3rd of my link. If I were to quote that
number, it would be embarrassing.

If you select your server well (use a Microsoft server, use
a download client opening 20 connections), you can get
tremendous source bandwidth. But, the result isn't
going to go faster than Speedtest.net results.

So just running Speedtest.net for 30 seconds or whatever,
is sufficient for "boasting purposes".

Servers vary, in terms of individual connection caps,
and the number of simultaneous connections they will allow.
For example, I was downloading some SDK from Microsoft, maybe
8GB in size, and Firefox said "24 hours remaining". That's probably
the slowest download I've ever been offered from Microsoft. I attempted
multiple times, to break and make the connection again, and could not
improve the rate. However, after I acquired a download client that
could open multiple connections, I did the math, figured out I needed
8 connections, and then the (double-sided) DVD downloaded in three hours,
instead of twenty four hours.

To make "rate-boasting" fair, it should be done with a
reproducible test case. And Speedtest.net is good enough,
considering the amount of time/effort these boasts are worth.

Paul
Ads