I concur with Travis, on what we'd like to see happen, but... I'd add that what 
he asks for may not be realistic or meaningful for MT.

First, remember Testing on Trango is bit different because it is a TDD based 
system. Its linktest gets run without ARQ utilized in the test. It also has the 
ability to have the test traffic override user data.  So it truly shows the 
loss/quality/capabilty of the link. Latency is ALWAYS consistent for each 
packet sent.  Its a total different game with CSMA/CA type protocol systems.

An 802.11 link shouldn't ever show packetloss (unless severally degrated) 
because of the re-transmission built into the 802.11 protocol, but has side 
effects of skyrocketing latency.

As well, the test should not be "TCP" based.  Having error 
correction/retransmission at layer3 also defeats the purpose of the test, and 
would not show the true packet loss of a link. 
Ultimately, it would be ideal for the test to be a layer2, but might be 
adequate if it was UDP or ICMP based.
Having retransmissions functionality in the test of any type, would always 
result in no packet loss, and not show the impact of sporatic latency that 
would be a side effect of a poor quality link.

Its very tough to do an accurate packetloss test, with other data on the link 
at the same time. And very hard to prevent other data from passing on a router 
type system that can be customized many different ways.

linkquality in Time based systems, show  by packetloss.
linkquality in Wifi based systems, show by varying speed.
 
So what can be done about it?

One possibilty is in additional to overall packetloss, show percentage of 
802.11 packets (layer2) from test data that were retransmitted at layer2 
(underlying packetloss).  

When we do tests, we do not jsut want to know whether a link is losing packets, 
we want to know if the ,link is performing optimal and if action should be 
taken to cure it.
For example, THis weekend we had a case where a link that normally works at 
54mb modulation, got misaligned and dropped down to 6mb modulation. The link 
tested out NO packetloss and high link quality, but it was only running at 
6mbps!!! If one has adaptive modulation, one needs to consider how to deal with 
that in the tests.

One requirement is to be able to set in the test parameters, what 
modulation/speed to test at.  So the modulation stays fixed during the test.

Another option is to have more advanced "latency" figures. For example, pings 
just shows average latency and peak latency. But what percentage of the pings 
were severally above the average latency?
So after the 1000 packets sent, it could report... average latency, peak 
latency, but also percentage of packets above 20% above the average, and 
percentage of packets 100% above average latency, percentage of packets above 
200% average latency.  Or something to that nature.   The idea being, what 
percentage of packets are not being sent optimally.   

The only true way to solve this, is to write a test that breaks the native 
802.11 mode, and runs the test at the driver level.

I guess what I'm saying is... It would nice to see it in the results output 
listed by Travis. But it may require insight from the MT programmers on what 
exactly is possible, and what method would be realistic to perform the test.

Tom DeReggi
RapidDSL & Wireless, Inc
IntAirNet- Fixed Wireless Broadband


  ----- Original Message ----- 
  From: Travis Johnson 
  To: WISPA General List 
  Sent: Friday, February 15, 2008 3:41 PM
  Subject: Re: [WISPA] MT tools


  Hi,

  I thought I had already posted what I would like to see:

  100 packets passed each direction, 10 times (1600 byte packets).

  Report back:

  Distance of link
  Error rate going from AP to CPE (in % as well as actual packet counts)
  Error rate going from CPE to AP (in % as well as actual packet counts)
  Throughput of the link

  Here is a sample test of a Trango AP to SU linktest:

  0  [AP Tx] 100 [AP Rx] 100 [AP RxErr] 0  [SU Tx] 100 [SU Rx] 98 [SU RxErr] 0  
504 ms  5079 Kbps
  1  [AP Tx] 100 [AP Rx] 100 [AP RxErr] 0  [SU Tx] 100 [SU Rx] 98 [SU RxErr] 0  
504 ms  5079 Kbps
  2  [AP Tx] 100 [AP Rx] 100 [AP RxErr] 0  [SU Tx] 100 [SU Rx] 100 [SU RxErr] 0 
 504 ms  5079 Kbps
  3  [AP Tx] 100 [AP Rx] 100 [AP RxErr] 0  [SU Tx] 100 [SU Rx] 98 [SU RxErr] 0  
504 ms  5079 Kbps
  4  [AP Tx] 100 [AP Rx] 100 [AP RxErr] 0  [SU Tx] 100 [SU Rx] 99 [SU RxErr] 0  
504 ms  5079 Kbps
  5  [AP Tx] 100 [AP Rx] 100 [AP RxErr] 0  [SU Tx] 100 [SU Rx] 99 [SU RxErr] 0  
504 ms  5079 Kbps
  6  [AP Tx] 100 [AP Rx] 100 [AP RxErr] 0  [SU Tx] 100 [SU Rx] 100 [SU RxErr] 0 
 504 ms  5079 Kbps
  7  [AP Tx] 100 [AP Rx] 100 [AP RxErr] 0  [SU Tx] 100 [SU Rx] 99 [SU RxErr] 0  
504 ms  5079 Kbps
  8  [AP Tx] 100 [AP Rx] 100 [AP RxErr] 0  [SU Tx] 100 [SU Rx] 99 [SU RxErr] 0  
504 ms  5079 Kbps
  9  [AP Tx] 100 [AP Rx] 100 [AP RxErr] 0  [SU Tx] 100 [SU Rx] 98 [SU RxErr] 0  
504 ms  5079 Kbps

  [AP Total nTx]    1000 pkts
  [AP Total nRx]    1000 pkts
  [AP Total nRxErr] 0 pkts

  [SU Total nTx]    1000 pkts
  [SU Total nRx]    988 pkts
  [SU Total nRxErr] 0 pkts

  [AP to SU Error Rate] 1.20 %
  [SU to AP Error Rate] 0.00 %

  [Avg of Throughput]   5079 Kbps

  Something like this (but in a graphical table inside of winbox) would be 
perfect. :)

  Travis
  Microserv

  Dennis Burgess - Link Techs Inc wrote: 
I work very closely with Mikrotik and I do understand what you are saying.
Lets do this...

What is the exact feature and results that we are looking for.

The reason why I ask, is they will ask!  

This is what I got so far:

10 second test
Test reliability of link
Overall throughput of the link
Packet loss, if any, on the link
Overall link quality

All of this calculated into ?  a Percent?   A number?  

Dennis M. Burgess
Mikrotik Certified Consultant
Link Technologies, Inc., St. Louis, Missouri
--WISP/Network Support Services--
+1 314-686-1302


-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On
Behalf Of Tom DeReggi
Sent: Friday, February 15, 2008 1:49 PM
To: WISPA General List
Subject: Re: [WISPA] MT tools

I guess the idea is that MT might be monitoring these lists.
But you are right, the best approach might be to Email MT directly.
But the truth is, Emailing support by an individual will likely not get 
results.
Manufacturers tend to want to see numerous individuals interested in the 
command or feature to jsutify the effort doing it.
By discussing these flaws on a list, is to get all the members familiar with

why these tools are needed, and possibly more members will add additional 
support inquiriies to request these improvements.

For one, I'd like to see the Mikrotik Consultants get involved in asking for

these features. If the consultants that specialize in the product don't 
understand what we need, its not that likely that the manufacurer would 
either. Consultants that represent the manufacturer have much more pull with

them, and a better relationship to discuss these things, and most 
importantly credabilty, to effect change.

With Trango, we effect change, and the reason is that we deal direct with 
the manufacturer and they hear us.  With channel support models, like 
Microtik offers, the manufacturer can lose touch with the end customer. It 
becomes an even worse problem when the manufacturer sells super cheap (which

we like), and the volume grows and each end user's opinion becomes less 
relevent.

One thing I'm very happy about was Mikrotik's support joining as a WISPA 
vendor member.  (Butch, possibly you one also ?)
I believe this gets WISP and Manufacturer closer in touch.

I'm hoping that these threads are not misinterpretted as manufacturer 
bashing, but interpretted as communication, that will effect product 
improvements.
The truth is MT probably offers better testing tools than most WIFI product 
out there.  But there is still room for improvements.

Tom DeReggi
RapidDSL & Wireless, Inc
IntAirNet- Fixed Wireless Broadband


----- Original Message ----- 
From: "Butch Evans" <[EMAIL PROTECTED]>
To: "WISPA General List" <[email protected]>
Sent: Thursday, February 14, 2008 11:59 AM
Subject: Re: [WISPA] MT tools


  On Wed, 13 Feb 2008, Travis Johnson wrote:

    What I really want is a way to right-click on an entry in the
Registration table and have an option that says "Linktest". It
would test sending 100 packets each direction, 10 times. It would
then report:
      have you emailed [EMAIL PROTECTED] with this suggestion?  I see
what you are wanting.  I suggested a way to get the data.  If that
isn't enough, then send an email to the folks that can do something
about it.  I hate these threads where the only thing that is
accomplished is a dead horse is mauled.

-- 
********************************************************************
*Butch Evans *Professional Network Consultation *
*Network Engineering *MikroTik RouterOS    *
*573-276-2879 *ImageStream                       *
*http://www.butchevans.com/ *StarOS and MORE                   *
*Mikrotik Certified Consultant *Wired or Wireless Networks        *
********************************************************************



    ----------------------------------------------------------------------------
----
  WISPA Wants You! Join today!
http://signup.wispa.org/

    ----------------------------------------------------------------------------
----
  WISPA Wireless List: [email protected]

Subscribe/Unsubscribe:
http://lists.wispa.org/mailman/listinfo/wireless

Archives: http://lists.wispa.org/pipermail/wireless/ 
    


----------------------------------------------------------------------------
----
WISPA Wants You! Join today!
http://signup.wispa.org/
----------------------------------------------------------------------------
----
 
WISPA Wireless List: [email protected]

Subscribe/Unsubscribe:
http://lists.wispa.org/mailman/listinfo/wireless

Archives: http://lists.wispa.org/pipermail/wireless/




--------------------------------------------------------------------------------
WISPA Wants You! Join today!
http://signup.wispa.org/
--------------------------------------------------------------------------------
 
WISPA Wireless List: [email protected]

Subscribe/Unsubscribe:
http://lists.wispa.org/mailman/listinfo/wireless

Archives: http://lists.wispa.org/pipermail/wireless/


  

------------------------------------------------------------------------------




  
--------------------------------------------------------------------------------
  WISPA Wants You! Join today!
  http://signup.wispa.org/
  
--------------------------------------------------------------------------------
   
  WISPA Wireless List: [email protected]

  Subscribe/Unsubscribe:
  http://lists.wispa.org/mailman/listinfo/wireless

  Archives: http://lists.wispa.org/pipermail/wireless/


--------------------------------------------------------------------------------
WISPA Wants You! Join today!
http://signup.wispa.org/
--------------------------------------------------------------------------------
 
WISPA Wireless List: [email protected]

Subscribe/Unsubscribe:
http://lists.wispa.org/mailman/listinfo/wireless

Archives: http://lists.wispa.org/pipermail/wireless/

Reply via email to