[nos-bbs] Interested in REPEAT packet statistics?

George [ham] VerDuin k8rra at ameritech.net
Fri Apr 16 14:33:25 EDT 2010


I accept that Kenneth.

On 04/16/2010 01:21 AM, Kenneth C Coe Jr wrote:
>  From a previous email: You have repeatedly represented yourself as a network professional
I was un-aware you read my material in that manner:
I am an EE by education.
The State of Mi has bestowed a PE licence so I can practice my trade.
My client base has been utilities and large business; the focus: power 
delivery and system protection.
I'm published, and had the emotional high of others standing on my 
shoulders to advance my work while I stand on the shoulders of others.
My second career required me to admin unix and design automation 
machinery while delivering automotive components.
I've even had fun implementing Grand Haven Musical Fountain [google it 
if you care].

So I'm not a network pro -- just someone with IT for 44 years of 
building, using, selling, software and doing OS stuff.
The software I designed to support my needs to provide client solutions 
produced asymmetrical analysis using symmetrical power system models, 
transient analysis on EHV DC transmission lines, pulling wires thru 
nuclear power plant construction zones, and machine actions from 
multi-computer control systems.
Networking is just something I do in support of the tools I create and 
use -- not my principal focus.
Therefore, while I've read the RFCs necessary to do the job at hand, I 
have no library of them.

I'll wade thru your use of network-specific details that you presume I 
understand, using research tools at my disposal, but let's establish 
English is my first language of choice, not the monikers of network minutia.

>
> Of course there are re-transmits. there is not a network in the world
> that doesn't have them. It is written into the standards because it
> cannot be eliminated. The question is never if there are re-transmits,
> but whether there is an unreasonable number of them given the status of
> the network.
If we are past chest thumping, here is proof you understand the point of 
my focus.
   "unreasonable"
It is my contention that the stats I presented demonstrate unreasonable 
repeats under VC configuration at K8RRA.
No more.
No less.
With detail to demonstrate network status.
Just one number that raises the red flag -- the idiot light -- saying 
"Look deeper".

> There is no answer to the re-transmits in your data, and
> you make no attempt whatsoever to provide a baseline of what is going on
> at the time of your tests.
>    
Actually the baseline is shown by the same stats under datagram 
configuration.
It demonstrates a reasonable success.

What I presented has not been seen before -- perhaps never again.
What I asked doesn't depend on performance of my node, it depends on 
years of experience.
The answer to my query centers on what you call unreasonable.

It certainly appears you don't understand the arithmetic I did for K8RRA 
node.
OK, I'll detail any number you choose to ask about.
Just don't jump to explaining why my radio is failing -- it isn't.

>
>> SO -- On the *single* issue of "Does a problem exist"
>>      
> Maybe, maybe not...
>    
I accept that assessment.
I don't accept that you have an analysis to present.
And I don't accept your expectation that I presented adequate data for 
you to analyse.

>    
>> Not how big is it
>> Not where is it
>> Not what caused it
>> Especially not what to do about it.
>> What do you say?
>>      
> If you go to a mechanic and say "my car is slow" he might ask if you
> have tried hitting the gas. The numbers you gave us can say nothing, and
> indicate nothing. there is no answer there, positive or negative.
>    
Good point.
Did you ask that clarifying question about the gas?
If you did -- I quite frankly miss it.
Or, did you jump directly to stupid?
HHMMM

>
> Move on only after you have completed this first step, but I sense that
> you may doing some work at this level. In the mean time, provide a link
> to download the dump, and we'll try to take a look for you.
>    
I have a spot to post the dump, I'm not convinced it is necessary yet.

You see I started this quest months ago after being ejected from a 
CONVerse session.
I found 32 repeat packets stacked up in a queue that prevented me from 
responding in a timely manner.
Simply 32 saturated my link to the outside world.
Not many packets in a global view of all packets.
Not perhaps statistically significant.
I'd happily give you more detail of this incident, but detail is 
probably not needed either.

At this moment, I'm only interested in the idiot light -- an idiom taken 
from earlier posts.
I'm interested in "unreasonable" -- and only in the context of a 
normally operating network.
I'm really NOT interested in having you determine the fix for what ails 
my node at this moment.
Perhaps we might reach that some day later, not today.

SO -- if you care to read my stats in detail, ask away.
Let's start at the top and drill down.
Lets start by deciding 29.5% is unreasonable [or not(?)] when 4.5% is 
the norm [see the original stats].

Yes on my terms -- I'm not asking you to doctor my system.
I'm asking for a consult on unreasonable.
And don't presume I already presented what you need to know in order to 
answer.

> Ken
>    
Skip




More information about the nos-bbs mailing list