Topic Options
#283 - 09/22/00 06:13 PM New feature feedback.
Pete Ness Offline



Registered: 08/30/99
Posts: 1106
Loc: Boise, Idaho
I'm working on prototyping some new features, and would like some ideas on how to incorporate packet loss into the main graphs.<br><br>My first attempt is here:<br><br><br><br>Notice the packet loss lines in the upper graph (the scale would be manually configureable - and it's set at 30% here). I'd love to get some feedback on this - and also some ideas on how to incorporate packet loss % in the lower graphs.<br><br>Thanks!<br><br>

Top
#284 - 10/01/00 04:04 PM Re: New feature feedback. [Re: Pete Ness]
LukeDukeNY Offline


Registered: 09/03/00
Posts: 4
Looks cool Pete.. as for your question on how to incorporate PL% into the lower graphs... what about using a configurable Color Gradient to show relative Packet loss.<br><br>i.e.- as packet loss increases, the colors could fade from black to bright red-- black being 0% loss, full red being 100% (or whatever scale the user chooses)<br><br>Just a thought...<br><br>

Top
#285 - 10/12/00 03:52 AM Re: New feature feedback. [Re: Pete Ness]
Anonymous
Unregistered


Why don't you add an additional graph for pl instead of interfering whith the response time graph?<br><br>U are doing a very good job, go on thanks!<br><br>

Top
#286 - 10/17/00 01:11 AM Re: New feature feedback. [Re: Pete Ness]
Pete Ness Offline



Registered: 08/30/99
Posts: 1106
Loc: Boise, Idaho
Here's a prototype I'm currently working with on the lower graph. I've got some other versions of this as well, but I like this one best so far.<br><br><br><br>I've also got another version of this where it uses the old bar graph style on the lower area, with the packet loss super-imposed. A bit more tweaking and I may post a picture of that too, for opinions.<br>

Top
#287 - 03/10/01 07:30 AM Re: New feature feedback. [Re: Pete Ness]
edcrowe Offline


Registered: 03/03/01
Posts: 9
Pete, in the image posted here what is the meaning attached to the height of the packet loss bars?<br><br>

Top
#288 - 03/10/01 01:41 PM Re: New feature feedback. [Re: edcrowe]
Pete Ness Offline



Registered: 08/30/99
Posts: 1106
Loc: Boise, Idaho
The height of the red is percentage of packet loss. The "Scale" of it is the red "50%" on the right side of the graph.<br><br>

Top
#289 - 03/10/01 02:23 PM Re: New feature feedback. [Re: Pete Ness]
Anonymous
Unregistered


i think you should kill the graph completely. the graph takes up to much space. just use the numeric values in the pl% column, ie. 10.25% or 15.75%. ping plotter is great because it is a simple and light on memory and processor resources as compaired with noetrace. neotrace is just a damn memory and processor pig with my dell xpsr400. i dont need a lot of jazzy features just quick and simple data on the current conditions of at@t worldnet in new jersey.<br><br>

Top
#290 - 03/14/01 05:33 AM Re: New feature feedback. [Re: Pete Ness]
Anonymous
Unregistered


I may be dumb here, but I can not make sense of this.<br>Percentage of what exactly?<br><br>Suppose I count apples, and keep a tally of % with worms.<br>Suppose I have counted plenty, and found a few percent with worms.<br>Now I count a one more, and it has a worm.<br>The % of bad apples CANNOT suddenly jump the way your graph does??????<br>I would have to have an awful lot of wormy apples to (gradually) climb to 50%.<br>Then I would have to have a long unbroken run of good ones to (gradually) reduce the fraction back down.<br>Confused,<br>Hugo<br><br>

Top
#291 - 03/14/01 09:50 AM Re: New feature feedback.
Pete Ness Offline



Registered: 08/30/99
Posts: 1106
Loc: Boise, Idaho
The percentage is based on the time period represented by one pixel width. If your lower graph is showing only 10 minutes, then each pixel is likely only representing one sample. In that case, is that one sample was a timeout, the red scale will be at 100% (just for that one sample). If, however, your graph scale is 24 hours, then each pixel could represent 15 or 20 samples (this is very dependant on your trace interval, however). In that case, 1 lost sample in that period would only represent 5% packet loss (if there are 20 samples in the period).<br><br>If your lower graph time scale is small, then a lost packet will likely represent 100% packet loss for that sample. As Ping Plotter has to average multiple packets into a single pixel width, though, there is a possiblity to have something besides 0 or 100% packet loss in a period. Here's an example of this where the % really shows a lot of information:<br><br><br><br><br> <br><br>

Top

Search

Who's Online
0 registered (), 13 Guests and 1 Spider online.
Key: Admin, Global Mod, Mod