Topic Options
#409 - 01/24/01 07:17 PM Ping spikes on even / odd hop numbers
Sappy Offline


Registered: 11/04/00
Posts: 2
I'll just make this simple by linking a nice big <A HREF="http://www.geocities.com/newbie61/pics/cdm-110-85.bent.tcac.net.png" target="_new">screenshot</A> of Ping Plotter as well a zip of the <A HREF="http://members.nbci.com/danewbie/1/cdm-110-85.bent.tcac.net.zip" target="_new">data file</A> (964KB), so that you can look at it however you want (I spent some time viewing one sample at a time (and noticed that the even/odd hops don't all ping high at quite the same time) - as the line graph does a pretty nice job of showing what's going on here). <br><br>If you have any idea of what might have been going on here, I'd sure like to hear it (as I'm dumbfounded about what could cause this); to me, it seems like it may have been something funny with PacBell, as the ping was spiking (by a whole 35ms or so ) on every even hop each 15 minutes (the odd number hops were the same... minus one occurrence, and offset by 7 minutes...) for 2 hours total.<br><br>[I have Ping Plotter 2.21 beta 3 running on a Win2K machine (used for solitaire and some browsing), which is a client to another Win2K machine (brother's - and we don't believe that he was doing anything which might cause this) sharing the DSL via ICS. The interval is set to 4 seconds, with a 250ms delay between hops, and a 16 byte packet size.]<br><br>

Top
#410 - 01/25/01 12:52 AM Re: Ping spikes on even / odd hop numbers [Re: Sappy]
Pete Ness Offline



Registered: 08/30/99
Posts: 1106
Loc: Boise, Idaho
Hi.<br><br>While this isn't the first time I've seen this, I've yet to come up with any very good reasons for this happening. Here's some pictures I've created from my cable modem (since replaced by DSL) that showed some similar "cyclical" tendancies. I, also, was doing nothing on my network at the times these graphs were recorded - and had this behavior for consecutive days constantly in several cases.<br><br>Now the interesting thing about your information is that it doesn't seem to directly translate into the final destination (notice how hop 12 on your graph seems relatively (but not entirely) unaffected by the problem). If you're looking to gather good ammunition, I'd try to find a time period where the "spikes" are visible in the final destination as well as some of the intermediate hops. If I were a network engineer confronted with data, I would be most concerned first with those issues where the end destination was noticably affected - as that's really the only thing that matters (heck, if all the hops through N-1 have 95% packet loss, but hop N has 0% packet loss and acceptable latency, then there's not really a problem).<br><br>If anyone else has any input on this "cyclical" problem, I'd be most interested to hear it. I pursued this with my ISP on 2 separate occasions with absolutely no resolution.<br><br>Sorry for the wide images, but it's easier to read that way...<br><br><br><br><br><br><br><br>

Top

Search

Who's Online
0 registered (), 8 Guests and 0 Spiders online.
Key: Admin, Global Mod, Mod