PING ponderings

[fiat_lux]

For the past 3 years, I've been telling newbies a ping is, in layman's terms,
"a measurement that shows the time, in seconds, it's taking the person you're
pinging to see what youre typing." But something occured to me tonight. When
you ping someone, your client makes a timestamp of the when the ping was sent
out, then it hits the other user, who's client sends a response back. When
your client recieves it, it makes note of the time and uses the difference
between it, and the timestamp from before to determine the ping time. But
this is really the time it takes to send something to the other user, and
have them send something back, or roughly double the time it takes them to
see what you're typing (my old ping defintion). So should I add "times 2" to
my definition, or do IRC clients cut this response time in half before
displaying it, to approximate the time of just sending a message from person
A to person B, instead of a round trip? I dont think it gets halved though;
my client's help file says "...to determine the length of time it takes to
send a signal to another person on IRC and then back again." So all this time
[So&So Ping response]: 4 seconds ..really means they see what I type in 2
secs?

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own

[Jack Lawson]

fiat_lux <dab...@geocities.com> wrote in message
news:7d7jg7$ckn$1...@nnrp1.dejanews.com...

I don't think so. I know how it works, but I'm not exactly sure how it does
work..if that makes sense. What it does is sends a privmsg to the
person/channel in the form PING 922182483 and then the person will reply,
with notice, the same thing. Then you subtract the time it was then, from
the time it is now. What confuses me really, is that if I'm in horrible lag,
and I ping someone, when I "pull out of it", if the person isn't lagging the
person still gets a reply of 0-1 seconds. There must be something that
compensates for it. Anyways, hope you found this helpful.

-WeaZeL

[James Atkinson]

You've pretty much got it right....a ping (the networking version thereof,
not the CTCP version..but I'm sure they are similar) is an ICMP echo
request. Computer A sends a ping packet to Computer B and if computer B
recives it it echos it back, Computer A recives it and checks the times and
determines the differance (mesured in milliseconds not seconds, if it takes
seconds for your packets to get to their destination and back you have
seriouse latency problems) and displays that, if a packet never comes back
Computer A chaulks (odd..I don't know how to spell that word...) that up to
'packetloss'. I don't know exactly how a CTCP ping works but I assume that
its about the same way, execpt mIRC rounds the millisecond responce times
out to human readable seconds, if its less then 1000ms(?) it just says 0
seconds...I hope this helps clear some things up :)

- James Atkinson
wil...@millenia.org

<insert disclaimer saying that all this might be wrong and its not my fault
here>

fiat_lux wrote in message <7d7jg7$ckn$1...@nnrp1.dejanews.com>...

>For the past 3 years, I've been telling newbies a ping is, in layman's
terms,
>"a measurement that shows the time, in seconds, it's taking the person
you're
>pinging to see what youre typing." But something occured to me tonight.
When
>you ping someone, your client makes a timestamp of the when the ping was
sent
>out, then it hits the other user, who's client sends a response back. When
>your client recieves it, it makes note of the time and uses the difference
>between it, and the timestamp from before to determine the ping time. But
>this is really the time it takes to send something to the other user, and
>have them send something back, or roughly double the time it takes them to
>see what you're typing (my old ping defintion). So should I add "times 2"
to
>my definition, or do IRC clients cut this response time in half before
>displaying it, to approximate the time of just sending a message from
person
>A to person B, instead of a round trip? I dont think it gets halved though;
>my client's help file says "...to determine the length of time it takes to
>send a signal to another person on IRC and then back again." So all this
time
>[So&So Ping response]: 4 seconds ..really means they see what I type in 2
>secs?
>

[Roel Spapens]

AFAIK mIRC (and other clients) just looks at the argument passed along in the
CTCP and as that is rounded to seconds it can't go under 1 sec, so the
rounding is done by how ping is implemented :)

I in ViRC simply scripted a more accurate ping, it sends out the ping and
notes the current time in variables (in ms), when the ctcp comes back I got
the actual time it took rather than the time that the ctcp reply brings back
with it.

In article <7d7oj2$pb...@mars.online.uleth.ca>,

[Jesse McGrew]

Roel Spapens <tallm...@yahoo.com> wrote:
[snip]
: I in ViRC simply scripted a more accurate ping, it sends out the ping and

: notes the current time in variables (in ms), when the ctcp comes back I got
: the actual time it took rather than the time that the ctcp reply brings back
: with it.

ViRC sends the current uptime in milliseconds ($mtime()) when you do /ping,
so all you need to replace is $decodepinginterval(). ;)

--
/ Jesse "Monolith" McGrew \ :) -> TMBG, UCB/TDS, AMD, ROTT/Sin
| Most quotable man on the Internet | :( -> CW/R&B/rap/alt, WWF/WCW/etc,
\ Mr2001 on IRC / Intel/Cyrix/PPC, Q2

[J Sneeringer]

On Tue, 23 Mar 1999 08:31:03 GMT, dab...@geocities.com wrote:
> I dont think it gets halved though; my client's help file says "...to
> determine the length of time it takes to send a signal to another person
> on IRC and then back again."

This is correct. The `ping' time you get is the round-trip time (RTT).
There would be little point in halving the time to get the the me-to-you
time, because there's no guarantee that the upstream and downstream data
rates are the same (e.g. PCM modems, DirecPC, ADSL), and no guarantee
that network congestion won't change between the request and the reply.

-James