Why do I get such different results for tracert and pingplotter. This is both typical and reproducible.<br><br>Using tracert:<br><br>Tracing route to NETCOMMANDER.COM [24.143.0.46]<br>over a maximum of 30 hops:<br><br> 1 14 ms 13 ms <10 ms 24.143.9.1<br> 2 13 ms <10 ms 371 ms 24.143.8.1<br> 3 41 ms 219 ms 234 ms 157.130.74.169<br> 4 27 ms 289 ms 233 ms 146.188.240.158<br> 5 28 ms 260 ms 220 ms 152.63.99.154<br> 6 68 ms 220 ms 220 ms 152.63.1.189<br> 7 96 ms 219 ms 220 ms 152.63.65.97<br> 8 2513 ms 68 ms 179 ms 152.63.69.145<br> 9 110 ms 96 ms 274 ms 157.130.110.66<br> 10 82 ms 97 ms 68 ms 24.143.0.46<br><br>From PingPlotter:<br><br>1, -------------- ,24.143.9.1,13,12,13<br>2, -------------- ,24.143.8.1,14,13,14<br>3,643.hssi3-0-0.gw1.hou1.alter.net,157.130.74.169,380,343,371<br>4,113.atm2-0.xr2.hou4.alter.net,146.188.240.158,769,655,646<br>5,192.at-1-0-0.tr2.hou7.alter.net,152.63.99.154,1099,999,1043<br>6,132.at-6-1-0.tr2.chi2.alter.net,152.63.1.189,1496,1304,1319<br>7,196.atm6-0.xr2.chi6.alter.net,152.63.65.97,1848,1607,1671<br>8,190.atm6-0.gw3.det1.alter.net,152.63.69.145,2211,1936,2002<br>9,netcommander-gw.customer.alter.net,157.130.110.66,2582,2282,2266<br>10,POSTOFFICE,24.143.0.46,2950,2590,2567<br><br><br><br><br>

Actually, I see signs of the same stuff in the TraceRT output (notice the first sample in hop 8 - it's 2513). The problem may be the same - just the TraceRT doesn't do it concurrently enough to capture a good picture of the problem. Do you get wildly swinging latencies in Ping Plotter also (where you'll sometimes see 200 ms latency on a hop, and the next samples see 2500)?<br><br>It is possible that the router at hop 3 has a problem with multiple outstanding ICMP requests. TraceRT only does one sample at a time, while Ping Plotter sends out up to 35 or so simultaneously. There are multiple possible reasons for this, here are a few:<br><br>* It's possible that your packet size is set to something that saturates your connection pretty quickly. I've seen this happen on nominal bandwidth connections where the packet size was large. Check in Advanced Options, Packet Options, Packet size and make sure this is something smaller than 100 (preferably between 10 and 56). <br><br>* A router in your system may not behave well when there are multiple outstanding ICMP requests. Testing this isn't too terribly difficult. First off, go to the "Advanced Options", "Packet Options" tab. From here, change the "Time interval between hop traces" to something in excess of your maximum expected latency (in the case of what you sent me, I'd use 400 or so). This slows down the rate at which samples are sent out. Now, change the "Trace Interval" in the main screen to something at least 20 seconds (this gives all 35 hops sent out time to complete before the next sample set it collected). Run your trace and see what happens. If your packet size (see previous point) is set to a reasonable number (10-56), then you should see the same results that TraceRT exhibits. Usually, a software update on your router can fix this problem.<br><br>The wild swings in latency seen at hop 3 (and beyond) even in TraceRT are quite odd. Keep in mind that Ping Plotter sends out all hops at once, where TraceRT does them much more serially (and much slower). Problem situations are pointed out much more quickly with Ping Plotter.<br><br>Please post back here what you figure out - or if you have comments on this analsys. A picture would be great as well - or send me a save file (data) and I'll do a bit more analysis on it.<br><br>

Thanks!!!!!<br>If I knew how to post the image I would show you before (defaults size56/25ms) and after (size10/400ms). Everything is pinging reasonable now.<br><br>I have recently had severe latency problems and swings with my favorite internet game. Could similar data transmission parameters be a cause?<br><br>

Actually, small changes in these parameters (the change from 56 to 10 is a small change) should affect your latency in a small way only (almost imperceptibly). If you're seeing wild swings in latency when making these changes, then there's likely something "not quite right" in one of the routers between you and the destination (probably one very close to where you're seeing the problem originate).<br><br>It really looks to me like the problem is NOT your parameters - but my picture from this end is very, very limited and I don't have all the details.<br><br>

Actually since I was implementing parameter changes in the same order as you gave your suggestions, I did get a chance to compare packet size impact. I would say the difference between size=56 and 10 was perhaps a 10% improvement. The real improvement I attributed to increasing time interval- tonight I shall try some intermediate values for this. Based on "a router's inability to handle multiple ICMP request" theory, should I expect a sharp break in ping times at some interval time value, or smooth relationship between interval time and ping time?<br><br>Also, tracert usually shows at least one hop out of the three sets being above 1500. My latency for some game traffic is very poor (often above 2000) and variable. Can you suggest any tests I can do to pinpoint the cause so as to better communicate with my ISP? They seem to want to help me but don't really know how. <br><br>

My interval time study results (30 sec avg, default packet size):<br>Interval,Ping Time<br>25,2770<br>50,2460<br>100,1970<br>150,1581<br>200,1171<br>250,760<br>300,385<br>350,108<br>400,86<br>Is this what should be expected?<br><br><br>

What kind of connection do you have to the internet? Just by the latencies, I'm guessing it's DSL/Cable, but I'd like a confirmation on this.<br><br>Generally, if you're dealing with a router that has a problem with multiple outstanding ICMP echo requests, you want to set it so that you don't have more than one outstanding at any one time. This is difficult to do from a timing perspective - as you can have bad responses at any time that will cause more than one to be outstanding.<br><br>Here's a procedure you can apply that will guarantee you never have more than one outstanding request - although it severely impacts performance.<br><br>Close all instances of Ping Plotter. <br>Open up your PingPlotter.INI file in your editor of choice (this file is in the Ping Plotter directory). <br>Insert a new line in the "Advanced" section of the .INI file. This line should read "MaxThreadCount=1" (without the quotes). <br>Open Ping Plotter. In the "Advanced Options", change the "Time interval between hop traces" to 0. (This ups the performance somewhat, isn't necessary, but very helpful). <br>Review the "Timeout speed" setting. If you never expect a hop to take more then 1 second to respond, enter 1000 here. Try to set it to the lowest reasonable number. In your case, you probably want to use 3000 - since you have bad latencies sometimes.<br>Turn *ON* "Use non-threaded Name Lookups". <br><br>This should get you the maximum performance without causing multiples to be outstanding.<br><br>The symptoms you're seeing aren't what I normally see when a router bios is misconfigured, though. Usually, I see really, really bad numbers, changing routes, and incorrect data - rather than just longer latencies. Let me know what hardware you have under your control (ie: routers, DSL modems, cable modems, etc.), and I'll see if I have heard of any problems with this.<br><br>