home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Ham Radio 2000
/
Ham Radio 2000.iso
/
ham2000
/
packet
/
terminal
/
pescav1d
/
pescador.eng
< prev
next >
Wrap
Text File
|
1996-05-19
|
38KB
|
937 lines
▒▒▒▒▒▒ ▒
▒ ▒ ▒
▒ ▒ ▒▒ ▒▒ ▒▒ ▒▒ ▒▒▒ ▒▒ ▒▒
▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒
▒▒▒▒▒▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒
▒ ▒▒▒▒ ▒▒▒ ▒ ▒▒▒ ▒ ▒ ▒ ▒ ▒
▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒
▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒ ▒
▒ ▒▒ ▒▒ ▒▒ ▒▒ ▒▒ ▒▒ ▒
Pescador Version 1.0 (Beta)
Beta Level #00148 19/05/96 (Expiration 07/96)
by Pedro E. Colla (LU1BUV) 1996
Packet Forward capture thru channel monitoring
(or ... The dog who is terribly at singing tangos!)
-------------------------------------------------------------------------------
Just in case it's not quite clear by now
THIS IS A BETA VERSION LEVEL
BETA BETA BETA BETA BETA BETA BETA BETA BETA BETA BETA BETA BETA BETA BETA BETA
-------------------------------------------------------------------------------
Introduction
------------
This program could get a forward out of the packet AX.25 net thru the
bare monitoring of the channel traffic, that is, without a physical connection
with a BBS, even without any transmision at all.
The operation is based on the capture of the forward between two any other
BBSs on the channel.
The methods and algorithms used are far from perfect, the scope and
limitations of the program will be reviewed in detail later on this document.
At the present stage the program implements a CONCEPT and this implementation
is yet to be refined to be trustworthy for the day to day operations; all
configuration options as well as it's algorithms are subject to change driven
by the field evaluation I receive or do myself.
Audience
--------
This program is targeted towards sysops and advanced stations who look
forward to improve the efficiency of our saturated packet channels.
It's assumed a basic familiarity with the radio amateur packet network
basic operational principles to being able to have some success with this
program.
If the program purpose sounds like chinese to you perhaps it's because this
program isn't for you.
About this document
-------------------
This document is a quick and dirty translation of the Spanish version
(PESCADOR.DOC) intended to ease the lives of the non-Spanish spoken
audience this program might have.
This is not the reference documentation for the program, the best update
effort is directed only to the Spanish version; further update of this
English version will be done on a 'best can do' basis.
Requirements
------------
- Hardware
PC compatible 386+, 640K RAM, some free disk space.
- Software
DesqView, Windows 3.1 u OS/2 2.1
Packet Switch G8BPQ 4.09a+ o TFPCX 2.10 (see configuration).
BBS o PMS enabled to work under the G8BPQ's switch and/or accept
a W0RLI formatted file import.
Mail in Packet Radio
--------------------
To fully understand the PURPOSE and SCOPE of this program is necessary
to understand the procedures who regulates the mail forward on the packet
network.
On a daily basis, the mail traffic flows thru the packet network using
multiple forward paths between the nodes (BBS) on the net.
Each forward session is implemented using a wide variety of links such
as HF, VHF, UHF, Satellites, Phone and even other networks.
Two information types are basically being transported, bulletins and
private mail; both widely differs in their handling.
The private mail jumps between any two net nodes following a route
defined by the destination of the mail; usually each net node will forward
the mail to the next step in the process based on the good or bad route
selection that this node sets for a given destination; usually too, the
private mail is cleaned from an intermediate node once it has been forwarded.
Bulletins, in the other hand, flows radially from the node who originates
it and without a predefined destination. To improve the definition of the
forward characteristics the bulletins are addressed to 'groups' such as
@LUNET or @WW; each group in turn has a 'topic' such as FBB@WW.
In general the group drives the forward path of a given bulletin and
the routing it will be under, but other considerations might apply at the
node level such as the individual size of the bulletin, the actual content
of it under the sysop judgement or the node who originates it.
When propataged, each node will forward a bulletin typically with all
the other nodes it has forward with, some selection based on the 'group'
name must be expected based on the target node forward profile.
As a difference with the private mail, the bulletin could be forwarded
several times (as many as other nodes forward with us) and would be left
resident on the node until some sort of maintenance purge process clean it.
Because of this difference, the bulletin taxes the net bandwidth
as a proportion of the nodes actually receiving it (plus the final users
who reads it); the traffic is then VERY SIGNIFICATIVE.
Purpose and Scope
-----------------
It's very important to understand the PURPOSE and SCOPE of this program
before to decide to use it.
The PURPOSE of this program is to take benefit out of the fact that in
crowded areas it's likely to found several forwarding nodes with good
radiolectrical visibility between them performing redundant mail forwarding.
In such cases the traffic flows thru the network and could be monitored
by the other stations simultaneously in the channel (althrough not activerly
performing forward themselves).
If thru this monitoring a fraction of the mail volume could be captured
then the forward will be avoided later.
To achieve this the program monitor the traffic between any two nodes
in pairs (up to 5 of those pairs are allowed) and when a mail forward between
them is detected as many messages of that forward are captured; the captured
messages are later formatted as an import file to be processed by the
local node (BBS/PMS) itself.
When the captured messages are, later, offered to the local BBS/PMS
they will be reported as already forwarded and the forward over the air
wouldn't take place.
The SCOPE of this program is to capture a big fraction as possible of
the whole mail traffic; this process isn't perfect and only a fraction
of the whole message flow will be captured.
It's important to understand some basic limitations the program will
show:
- The link being monitored operates as an AX.25 connected circuit,
both could use the protocol to overcome link errors; a pasive
monitoring station couldn't use those those methods. Link errors
at the monitoring stations not occurring at the receiving stations
at the same time (thus activating the protocol error recovery
procedures) will hit the whole piece of mail where the error
occurs.
- Exception given for trivial interceptions (i.e. monitoring ourselves)
there will be relative difference on how the monitored stations are
receiving themselves and how, at the same time, we receive both
stations. This effect is noticeable both in terms of signal strenght
or the impact of a colision with a 4th station operating in the
same channel.
- The widely used method of forward the messages as a compressed batch
(i.e. FBB) optimizes the link between the stations making the forward
but inherently makes the monitoring harder.
Usually some error during the reception of a batch will hit several
messages.
- At the same time, the partial transmision of messages (such as the
offset recovery method used by the FBB BBS) makes extremely hard
for those messages to be captured; in fact this program skips those
messages.
Some of the above factors are compensated by algorithms imbedded in the
program thru the usage of some 'fuzzy logic', the compensation isn't
complete though.
Due to those factors THIS PROGRAM MUST NOT BE UNDERSTOOD AS A REPLACEMENT
FOR THE CONVENTIONAL FORWARD, it has to be taken as a complement instead.
It's still required some considerable experimentation effort and fine
tunning over the program algorithms to improve the capture efficiency;
preliminar figures shows a capture efficienty between 10-60% on a W0RLI
type of forward (ASCII non-compressed) and 5-50% on a F6FBB (binary,
compressed) forward.
The above ranges assumes a good electrical visibility between the
monitoring station and both stations been monitored; figures could be
even higher under good conditions.
The criteria and program philosophy is to provide some savings at
a very low execution cost in resource terms; so, if a single message is
captured thru this way it could be considered a saving on future bandwidth
usage.
Operations
----------
To understand the program operation a reasonable understanding of the
AX.25 protocol is required.
The program monitors the traffic on the channels visibles at a node,
usually working under a G8BPQ NET/ROM switch altrough a posibility exists
to operate it directly with a TNC (ver TFPCX).
It's assumed that the program will run concurrently with some BBS or PMS,
and potentially other components sharing the switch; this makes mandatory to
run it on a multitasking environment. This requirement could only be avoided
if the program run exclusively on a node, but this scenario it's very
unlikely on practical terms.
Since the program runs on DOS and the G8BPQ switch provides support for
the industry most relevant multitasking environments it's likely to expect
it will operate satisfactory on many of them.
In quiescent state the program simply monitors the channel traffic without
reacting to it, except when the monitor of the traffic is activated.
To start the monitoring of a pair of nodes both the nodes address
(callsign plus SSID) and the forward method (W0RLI or F6FBB) must be
configured by the sysop; the program itself will not define that.
The monitor of the activity between the defined link pairs will continue
without effect until a mail 'proposal' consistent with the forward method
informed is detected. Usually that means a FA/FB command for F6FBB and
SP/SB for W0RLI.
Once triggered by the proposal the program will start to follow the
frame sequence being interchanged by both nodes under the AX.25 protocol
guidelines trying to intercept and interpret those frames that could be
validated as correct, from the intercepted frames only those with information
are relevant altrough all of them are verified with flow control purposes.
The resulting captured information, free of control codes, are stored
using an import format to be able to feed them into the local BBS/PMS later;
the used format is basically an ASCII text file delimited by the forward
command (SB/SP) and an end of message marker (/EX).
Once the message end is detected (or the sequence is lost) the program
fall back to monitor the traffic between both nodes trying to catch the next.
Not all the AX.25 protocol is followed; the program mimics a TNC only
partially, only relevant frames are decoded.
In case the sequence followed shows that some information were lost the
program activates algorithms to start a recovery process of the sequence;
if the action is successfully completed the message capture will continue
otherwise the message is completely lost.
As an example, if the sequence shows a frame has been missed the program
logic wouldn't interrupt the message capture until be sure the receiver node
for that frame didn't lost the same frame too; if the receiver had the same
problem (i.e. a collision) the receiver TNC will start the protocol
correction and retry handshake leading eventually to the missed frame to
be resent.
However, when a frame is positively lost the complete message is lost;
if the link is operating using a batch message forwarding process it's
also probable for the rest of the messages in the batch to have some problems;
the logic tries usually to recover from that, but the degree of success is
quite variable.
Monitoring Slots
----------------
The program is able to monitor up to 5 link pairs simultaneously using
each one their own transfer protocols (FBB o RLI).
Each channel or link pair is referred as a 'slot' through this document.
The stations associated with each slot and the forward method used by
them are configured on the file PESCADOR.CFG and could be temporarily
altered at execution time using the F7 key.
It's on the sysop care not to assign the same stations (callsign and SSID)
pairs to different slots (including using different protocols) because that
will drive unreliable and faulty behavior in the forward capture process
(some of the frames will be captured by one slot and others by the other).
Installation
------------
The distribution file has to be decompressed using the ARJ utility on a
directory where the program will be run (usually \PESCADOR), this isn't
mandatory.
It's assumed a multitasker (if used) and the G8BPQ switch (or TFPCX) are
already properly installed and working.
Then, to install the program the following steps has to be performed:
- Decompress the distribution file, the included files are:
PESCADOR.EXE Main executable program.
PESCALZH.EXE Auxiliar program.
PESCADOR.CFG Configuration Example (G8BPQ).
PESCA.BAT Execution batch example (G8BPQ).
PESCATFP.BAT Execution batch example (TFPCX).
PESCATFP.CFG Configuration example (TFPCX).
PESCADOR.INI TFPCX configuration example.
PESCADOR.DOC The Spanish version of the manual.
PESCADOR.ENG The English version of the manual.
README.1ST Release news.
- RTFM, specially the README.1ST.
- Configure the PESCADOR.CFG file, use the example as a guideline.
Mandatory parameters are
MYCALL STATION_A STATION_B
SLOT
- Define a DOS session where the program will run (if using a multitasker).
- Configure the PESCA.BAT adapting it to your machine environment.
- Adapt the PESCAFWD.BAT procedure (mail import) to your BBS/PMS.
- Launch the program using PESCA.BAT.
Configuration File (PESCADOR.CFG)
---------------------------------
Follows the description of the parameters available to customize the program
behaviour, some parameters are mandatory and some optionals.
Mandatory parameters doesn't usually have a default value, omission lead
to the program cancellation.
Some of the parameters values (both mandatory or optional) could be later
modified during execution using the available commands; however, as this
program will usually run unattended it's very important to pay due attention
to the configuration process.
INTERFACE
It's the interface the program will use to establish a dialog with the TNC.
Options are BPQ or TFPCX, the first is assumed as default.
Example:
INTERFACE BPQ
MYCALL (mandatory)
This is the callsign of the local station, without SSID.
Example:
MYCALL LU1BUV
HADDR (optional)
It is the hierarchical node address, default is none.
Example:
HADDR #ADR.BA.ARG.SOAM
Note: The activation of the R: record generation activated by the RLINE
parameter requires this one.
BPQ_PORT (optional).
This is the BPQ stream used by the program to monitor, default is stream 10.
This parameter is required only if working with the G8BPQ Switch.
Example:
BPQ_PORT 10
TFPCX_INTR (optional).
Configure the interruption number used by the system to communicate with
the TFPCX driver.
Default is 253 (0xFD).
This parameter is only required when using the program using the TFPCX
driver, IT MUST BE INFORMED AS A DECIMAL NUMBER.
Example:
TFPCX_INTR 251
SLOT (mandatory)
Configure the callsigns being monitored on a given slot and the forward
protocol used for that.
Up to 5 of these parameters could be used, defining each one a slot
numbered from 0 to 4.
Format is:
SLOT Slot#,Origin Station,Target Station,Forward Mode
Where:
Slot# Slot number [0..4]
Origin Station, Target Station Stations to monitor.
Forward Mode Protocolo [FBB..RLI]
The stations must be informed with SSID, that is, as they show
monitoring the channel.
If the forward takes place using a NET/ROM circuit the L4 callsign
must be indicated and not the L2 callsign.
This instruction set doesn't have defaults, if not configured no
monitoring will take place.
If a slot is informed more than once the last informed will be
retained.
Example:
Slot 1,LU9DGN,LU5DWT,FBB
LOG (Optional)
Activates the monitoring activity to be reflected on the logging
file named PESCADOR.LOG.
The logging will reflect the content of the online monitoring activated.
Default is OFF (no logging).
Example:
LOG ON
LOGFILE (Optional)
This parameter configures the file where the logging will take place
if activated.
As a default this file is named PESCADOR.LOG; it could be useful
to assign a different name if more than one instance of the program
is run at the same time.
Example:
LOGFILE C:\PRUEBA.LOG
MONITOR (Optional)
Configure the starting monitoring detail level, the detail will vary
between level 0 (no monitoring), 1 (headers only) and 2 (headers+info).
Example:
MONITOR 0
TRACE (Optional)
Configure the trace detail level, the detail will vary between
level 0 (no trace), 1 (logic flow) or 2 (flow+info).
This option is only useful for debugging purposes.
Example:
TRACE 1
PRIVATEMAIL (Optional)
Configure if private messages are exported, default is ON (yes).
Example:
PRIVATEMAIL OFF
RLINE (Optional)
Configures if an R: record has to be added to captured messages as
if captured thru conventional forward.
Default is OFF, if activated requires the configuration of the
HADDR parameter.
Example:
RLINE ON
EXPORTTIME (Optional)
Configures the elapsed, in minutes, between the sucessive executions
of the export procedure (PESCAFWD.BAT).
Example:
EXPORT 60
EXPORTFILE (Optional)
Configures the path and name of the file where the messages will be
exported.
The default name is PESCADOR.FWD and will reside at the same path
the program is executed from.
Example:
EXPORTFILE \TSTHOST\PESCADOR.IN
VIDEO (Optional)
If difficulties are experienced with the program display, usually
shown as a black or empty screen under an otherwise normal operation,
it could be produced by a video mode conflict.
As a default a color adapter is assumed (VIDEO COLOR), if irregular
behaviour is detected try VIDEO MONO.
Example:
VIDEO MONO
ERRORMAIL (Optional)
This parameter configure is marginal mail captured with uncertainity
about it's integrity has to be captured or not.
This parameter has only trascendence capturing a batch forward
protocol such as the FBB where a hit on a single mail could destroy
the whole batch capture.
The default value is active (ON), see the 'Mail Marginal' section for
operation details.
Example:
ERRORMAIL OFF
ERRORFILE (Optional)
Configures where the marginal mail intercepted when the ERRORMAIL
parameter is active is stored.
Default value is PESCADOR.ERR at the same path where the program is
executed from.
Example:
ERRORFILE PRUEBA.ERR
PESCA.BAT File
--------------
This batch file is an example on how the program could be started,
it's use is recommended and a customization to the local execution
environment is required.
PESCAFWD.BAT File
-----------------
This batch file is called internally by the program to transfer the
captured mail at regular intervals into the BBS/PMS automatically.
The regular intervals are governed by the parameter EXPORTTIME, if
0 this procedure isn't executed.
Configuration to the BBS/PMS used is required, the batch has to work
in a way that the content of the PESCADOR.FWD file is dumped into other file
where the BBS/PMS will take the export from.
Once dumped the PESCADOR.FWD file must be erased; the main program will
guarantee this file isn't in use when the external procedure is called.
Program Operation
-----------------
The program could be executed from the DOS prompt or from a batch file
such as the one provided as example (PESCA.BAT).
At the start the execution behaviour is dictated by the parameters
indicated as execution switches or by the actual content of the PESCADOR.CFG
file.
The format is:
PESCADOR [-c ConfigFileName | ? ]
Where:
-c ConfigFileName
This parameter could be used to specify a configuration file
different from the default (PESCADOR.CFG).
It could be useful if a test configuration is generated or if
several instances of the program are run at the same time.
If a configuration file isn't found the program aborts execution.
-? Execution Help.
A short help message is printed, the execution is stopped.
Execution
---------
This program is designed to be run on a multitasking environment as
a 'partner' of a BBS/PMS saving channel bandwidth on those messages
that could be captured by simply monitoring of the channel.
A particular case, covered later, when the program runs under TFPCX
is an exception to the above scenario.
The machine resources taxed by the program are modest; 150 to 200K of
RAM memory are occupied by the program. Probably a 256 to 384K window
should be enough.
IMPORTANT, if the program works normally but fails to export at
periodic intervals the captured mail it could be from a lack of memory
to launch the PESCAFWD.BAT procedure, try increasing the RAM memory
available to the program.
It's convenient to launch the program using a procedure such as the
provided PESCA.BAT with the proper customization to the local environment.
The batch PESCA.BAT will usually run thru the following steps:
1) DOS environment definition (such as PATHS or video modes).
2) Some sort of TAME program (which is optional, but convenient).
3) Load of FOOLBPQ (necessary when running G8BPQ under OS/2 only).
4) Default execution of PESCADOR.EXE.
5) End of execution, handling of abnormal terminations.
Each one knows what his machine has and how far it could be pushed with
additional load; in my case the program runs on a 386SX 16 Mhz Intel box
under OS/2 at the same time than TSTHOST and JNOS without a noticeable
degradation; on bigger/faster machines the operation should be even more
unnoticeable.
A reasonable multitasking environment couldn't be established on older
machines (such as ATs or XTs) so the program probably isn't well suited
for them in real world terms, it would run though.
OnLine Commands
---------------
At the execution start the program will use the configuration indicated
at the PESCADOR.CFG file to define it's operational behaviour.
Some configuration could be performed using online commands.
F1 - Help.
Shows a summary of the available commands.
F2 - Application Monitoring.
Defines the level of the monitoring being performed.
Sucessive activation lead to the level Off (no monitoring), On
(General monitoring) and Inf (Detailed monitoring).
Monitoring will be filed also at the log if enabled.
F3 - Channel Monitoring.
Defines the level of detail on the information at the AX.25 protocol
(channel activity) level.
Sucessive activation cycles the level between Off (No monitoring),
On (General or CallSign monitoring) and Inf (Detailed Monitoring).
Monitoring will be filed also at the log if enabled.
F4 - Export
Will execute the PESCAFWD.BAT inmediately.
This command will be effective if the program is such an internal
state that safely allows the PESCADOR.FWD file (or equivalent) to
be manipulated (and IF it exist).
F5 - Reset Sesion
Reset the slot individually to a idle condition.
F6 - Error Mode.
Enable or Disable the capture of marginal mail into an error file
(PESCADOR.ERR or equivalent).
F7 - Slot Edit.
Enable the editing of the slots (stations and forwarding protocol).
The modified slots will be reset prior to use.
Modifications thru this command has the lifespan of the current
execution because they are not reflected on the configuration file.
F9 - Session Logging.
Cycles the logging On and Off, each time the logging is cycled from
Off to On the existing file is also erased.
F10- DOS Gateway
Allows a DOS shell to be activated, while active the monitoring
is STOPPED.
a-Q End of Execution (Alt+Q).
Unconditionally ends the execution.
Use of the captured traffic.
----------------------------
As explained before the mail successfully captured is stored on the
PESCADOR.FWD file as an W0RLI import, this format is accepted widely
on most packet BBS as a way to import messages.
This file has to be, at regular intervals, feed into the BBS/PMS to
load the messages on it. This process could take place either regularly
or during some maintenance cycle of the BBS/PMS.
A balance has to be exercised on this load; too frequently will end up
interfering with the capture process itself.
Too sporadic and the mail will arrive thru normal forwarding before it
has the opportunity to be feed into the BBS/PMS; a balance has to be
judged by the sysop based on his operational pattern.
An import each hour seems to be a good balance on most cases.
Please note that the capture is basically a blind process, the same
message could be captured several times on different slots and even a
message already at the local BBS could be captured; however, it's assumed
that the BID control mechanisms on the BBS/PMS will filter those duplicates.
TFPCX Operation.
----------------
The program peak benefit is obtained operating it as a BBS/PMS partner.
That implies the simultaneous operation under a multitasking environment
and some means to share the hardware (ports and TNC), such as the G8BPQ switch.
There are scenarios where it might be apropriate to run the program off
from this concept, using a dedicated TNC and port as an example.
This would usually be a test scenario, such as when a quick evaluation
ir required without the complexity of a full installation and configuration.
To cover that, the capability to use TFPCX as a driver with the TNC is
provided; to install the program in this way proceed as follows:
1) Install the TFPCX Version 2.1 driver or later.
2) Configure the driver using PESCADOR.INI as a template.
3) Configure the program, you could use PESCATFP.CFG as a template.
4) Launch the program as indicated on the example PESCATFP.BAT.
Legal and Ethics.
-----------------
Some sysops could question the whole legal base for this program, after all
it's able to capture mail from other stations without their consent, even
without their knowledge.
Each country will have his legal scenario I am not able to talk over, each
sysop must seize it.
In my own country, which in many respects follow the trends set by the
FCC regulations, the usage is perfecly legal.
Since the system is completely pasive (it don't transmit) I think it's
even legal to be used by SWLs and amateurs without license!!
However, even closing the legal issues some sysops could raise some issues
regarding the ethics of the message interception.
From it's the author point of view that a bulletin is publicly being
transferred on a open network under a known capability of many other stations
to monitor that transfer; the bulletin originator itself has no knowledge
of the forward process of it's own material once it's on the net.
Private mails are, under the author's opinion, on a similar conceptual
framework. However, a more 'touchy' esence is recognized on this case.
If the sysop feels the capture of the private mail is in contradiction with
his/her policy it could easily avoid that capture with the PRIVATEMAIL OFF
configuration sentence.
Operation Notes
---------------
The program capture efficiency will vary depending on several factors, such
as:
- Quality of reception of the intercepted nodes with us.
- Quality of reception of the intercepted node between them.
- Channel activity and colission level.
- Activity being generated by our local BBS/PMS on the monitored channel.
- Message size.
The capture proficiency, expressed as messages successfully captured over
total messages interchanged on a given link, deteriorates very quickly with
a decreasing reception quality and with channel activity.
Also, the proficiency degrades very sharply as the average size of the
messages being forwarded increases; this is reasonable since as the message
length increase the probability of a hit on it increases too.
The capture is degradated by a local activity at the BBS/PMS on the
monitored channels; this is because our own BBS/PMS will prevail on
colissions with other stations, including the ones being monitored.
A reasonable expectation should be set on a capture in the order of 5 to
50% on FBB monitored links and somewhat higher on RLIs, this is because
on RLIs a single hit impacts a single message while on FBBs a single hit
could destroy several messages on the batch.
The efficiency will vary, tipically, with the hour. This is because as
peak activities occurs less messages will be captured, peak activity
occurs usually at given hours on a given channel.
However, a global proficiency increase could be achieved with the
simultaneous monitoring of several link pairs, specially if the forward
is common to all BBS on the area.
Thus, a message lost due to some problem on a link could be captured on
another; as a counterbalance it's very likely to capture a given message
twice.
Out of experimentation, the proficiency rises a 50% by each link
monitored, in other words it three nodes with low proficiency are monitored
such as:
Enlace A-B 9%
Enlace A-C 12%
Enlace C-H 8%
The overall proficiency will be in the order of 20%; this value increases
noticeably as soon as the monitored links increases the individual
proficiency, as an example:
Enlace A-B 15%
Enlace A-C 25%
Enlace C-H 18%
Will yield a total proficiency of about 40%, which is by all means
significative in terms of bandwidth savings because of lack of forward
(our node will look to the network as using only 60% of the usual bandwidth
for forwarding purposes!!!).
Dara una eficiencia total del orden del 40%, lo que es significativo en
terminos de forward adicional que se evita con nuestra estacion.
Estos guarismos seguramente variaran en otros escenarios de forward
distintos a los que se ha experimentado el programa hasta el momento.
As a general concept, consider this program as less proficient than your
own station performing connected transfers, if you experience problems
even on connected links (such as collisions, retries, hidden terminal
effects, etc) this program will be impacted even greatly because it lacks
the protection for fault recovery provided by the AX.25 protocol on
connected mode.
Further from this point no theory is available, just try it and bring up
your own conclusions!
Mail Marginal
-------------
Whenever a L2 AX.25 problem is detected (such as the complete lost of
a whole frame) the message being received is abandoned because it's
integrity couldn't be reasonably guaranteed (we don't have a way to
guess how important is the missed frame).
On a RLI protocol forward this message is gone and the program prepares
itself for the next one.
However, on FBB protocol forwards messages comes in batches, an error
during the reception of a message impairs the whole batch (worst case
is when the frame contains the end of a message and the start of the
next).
The program in such cases activates an algorithm that explore the binary
transfer itself and tries to sync with the start of a new message, this
is called SmartSync, and if the message is successfully captured it's
stored.
A similar case occurs when the lost frame is the answer of a proposal
(then we don't have the foggiest idea about the message sequence as
sent by the sender); the message itself could be flawlessy captured but
the program couldn't associate them with the apropriate header (then
the program don't know to who the message is sent, who originated it
and it's BID).
On both cases, and if the ERRORMAIL parameter is ON, the message is
exported on a special file named PESCADOR.ERR as a private mail
directed to the sysop and from the sysop.
Before being exported, the message itself is scanned, and all the
information that could be rebuilt from the message itself (such as
R: records or 'From' or 'To' sentences) will be used to reshape the
origin or destination of the mail as well as it's bid.
The quality of those messages is much lower than the ones guaranteed by
a flawlessy capture with all the guards of the AX.25 protocol so they aren't
imported automatically on the BBS/PMS by the PESCAFWD.BAT procedure; the
sysop has to decide if he wants to manually inspect or correct the message
and import them or just discard the whole marginal capture.
Final
-----
This section has a disproporcionate size with the description of the
program; but I have years teaching me that this program will contact
me with a majority of wonderful people and with few idiots, thefts and
crazy guys.
If you feel you belong to the first category please don't bother by
the following paragraphs, they're not for you; if you feel you could
qualify to the second perhaps you'll do a bright thing not using this
program in the first place.
The usual stuff, the program has been tested and works reasonably within
the documented behaviour, but any failure or damage caused by its operation
is beyond my responsibility. Use it at your own risk.
As often stated: 'The only thing I guarantee regarding this program it that
it will use some of your disk space'.
This program has been written for a radioamateur, for radioamateurs on
radioamateur uses; within those limits use the program as it best suit
your wishes for no charges or fees. However, I do retain the full
copyright on the program and you don't have any legal right on it.
You're expected to cease on the program utilization at the first formal
notification you receive, if you don't share this condition just don't
use the program.
Please, take into consideration this is a BETA version of the program
released with the purpose of an open evaluation and feedback.
I did invest a considerable amount of my own time in the study, design,
build and test of this program; I really enjoyed the experience (and
the learning) I got from it and I don't expect your thanks.
However, I do expect you to be kind and constructive on any comment you
dare to send to me on the program; if you're tempted to send other kind
of comment just don't waste your time nor mine.
It's IMPOSSIBLE to verify and test all possible hardware and software
configurations in operations out there, I don't discard conflicts between
this program and others, so if you report an error to me please give me
a reasonable description on the how, when, wheres of the environment
in which this program is intended to be run.
I usually do honor my mail reasonably fast, but be reasonable yourself
and please remember that this is a hobby to me (like it's to you) and
I might have other priorities on a given momment that could make it
difficult to me to answer mail on the expeditive way you might expect.
The support or any further update of this program could cease without
previous notification.
This program is free for radioamateur purposes, it's commercial use is
prohibited; it's distribution couldn't be charged except the nominal
value of the recipient, the effort to record it and any distribution
fee involved.
Comments and suggestions please directed to:
AX.25 : LU1BUV@LU1BUV.#ADR.BA.ARG.SOAM
AMPR : lu1buv@lu1buv.ampr.org
Fido : Pedro Colla @ 4:901/267
Internet: colla@pec.pccp.com.ar