home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Media Share 9
/
MEDIASHARE_09.ISO
/
hamradio
/
kant405e.zip
/
KANTNODE.DOC
< prev
next >
Wrap
Text File
|
1992-07-19
|
10KB
|
255 lines
G8BPQ Packet Switch for the Kantronics Data Engine.
I have now produced a version of my packet switching software for the
Kantronics Data Engine. The facilities provided are very similar to
those of the standard (PC) version, but there are slight differences,
mainly due to the different hardware environment. The two radio ports
are configured as if they were HDLC cards, but the async port may be
used either as a KISS (or NET/ROM) link to another TNC, or as a 'normal'
TNC link to a PC. Eventually this will provide all the options of my
Virtual TNCs (ie TNC2, PK HOST mode for AA4RE or UFQ), but so far only
the TNC2 version works. Obviously (or perhaps not so obvious really),
as there is only one serial port, you can only define one 'Virtual'
TNC in the TNCPORT params. Also the async port can only be used for
one thing at a time, but you can configure both, and select either
link mode or TNC mode using a switch. This is useful if you are setting
up a multiport node, but want to attach a terminal temporarily whilst
configuring the system. The SYSOP functions are also provided within
the code.
Environment.
The code normally runs from EPROM (one 27512), and I suggest that it
replaces the Kantronics EPROM. A facility has been designed to allow it
to go in the second socket, and be selected using the front panel switch,
but as I have mislaid the chip that Kantronics kindly sent me, I havnt
been able to test this feature. It is also possible to down line load the
code, and run it from RAM, which is useful when experimenting, but not
ideal for a live system. It also requires a 128K byte static ram chip,
which is rather expensive! The PROM based system requires 2 * 32k byte
static RAMs.
The program is distributed in a file KANT.BIN, which must be merged
with the BPQCFG.BIN file produced in the normal way, and with PASSWORD.BPQ,
to produce either a PROM Image (KANT.PRM) or an Intel HEX format file for
down line loading (KANT.HEX). This is done using program KMERGE.COM. Normally
a PROM image is produced - to get the Intel HEX output, type KMERGE HEX. Note
that the current HEX file format is not compatible with the loader in the
Kantronics standard ROM. It may well soon be but in the meantime you must
use the supplied KBOOT program (again blown into a 27512, but this time
in the top 16k of the prom) if you want to use the down line load option.
The SYSOP facility is password protected. The password string is taken from
file PASSWORD.BPQ. To enter SYSOP mode, enter PASSWORD. The system will
respond with 5 numbers. Then enter PASSWORD ABCDE where ABCDE are the
corresponding characters from the password string in the EPROM. The system
should then respond Ok. Note that the password is not case sensitive.
There are a few additional SYSOP commands:
CLOCK displays or sets the battery backed real time clock. To set the clock,
enter CLOCK HH MM. As the clock is only used for timestamping the ROUTES
table, only the hours and minutes registers are used.
COLDSTART reinitialises the RAM, and resets the node. Normally the ROUTES,
NODES and STATS are maintained when power is off.
WARMSTART restart the code without clearing the ROUTES, NODES and STATS.
A digi'ed link may be added to a route entry. Enter:
route call port v digi1 (digi2)
A locked node entry may also be added:
node add alias:call neighbour-call port quality
Or deleted:
node del alias:call
The following SYSOP commands are new in version 4.05e, mainly to save
having to keep blowing new EPROMS when experimenting.
Command Corresponding BPQCFG parameter
OBSINIT OBSINIT
OBSMIN OBSMIN
NODESINT NODESINTERVAL
L3TTL L3TIMETOLIVE
L4RETRIES L4RETRIES
L4TIMEOUT L4TIMEOUT
T3 T3
IDLETIME IDLETIME
LINKEDFLAG ENABLE_LINKED
IDINTERVAL IDINTERVAL
MINQUAL MINQUAL
FULLCTEXT FULL_CTEXT
HIDENODES HIDENODES
L4DELAY L4DELAY
L4WINDOW L4WINDOW
BTINTERVAL BTINTERVAL
Switches and LEDS.
The switch selects LINK or TNC mode on the ASYNC PORT. (In for LINK mode)
LED1. On in TNC mode, flashes as packets are received on async port
in LINK mode.
LED2. Packet Received on A port.
LED3. DCD on A port.
LED4. Not Used.
LED5. Packet received on B port.
LED6. DCD on B port.
LED7. CSMA Indicator. Shows 'Slottime Wait' timer is running
LED8. Flashes at one second intervals to show system is running
'Features' of current version (4.03)
Operation of the async port in Host Mode is not supported.
Configuration Suggestions.
The following are some ideas on how the Data Engine could be used in
packet networks. (Using my software - it can, of course be used as a
normal TNC with the Kantronics software).
Standalone Data Engine.
This would seem to be a reasonable solution for a small node,
providing one user access and one trunk channel. The main advantages
over a pair of conventional NET/ROM or TheNet TNCs are that only one
Call/Alias pair is required, so the size of NODES lists and broadcasts
are minimised, and the STATS facilites allow better network management
and performance monitoring. Also it will be better able to handle higher
speed links as they become available. The main disadvantage is a rather
higher cost.
Data Engine driving KISS TNCs.
----------------- ---------------
| |------------------| 'Normal' |-- Radio 3
| Data Engine | | |
| | | KISS TNC |
----------------- ---------------
| |
| |
Radio 1 Radio 2
----------------- ---------------
| |------------------| |-- Radio 3
| Data Engine | | KPC4 |
| | | |-- Radio 4
----------------- ---------------
| |
| |
Radio 1 Radio 2
----------------- ---------------
| |------------|-----| 'BPQKISS' |
| Data Engine | | | |-- Radio 3
| | | | TNC |
----------------- | ---------------
| | |
| | |
Radio 1 Radio 2 | ---------------
|-----| 'BPQKISS' |
| | |-- Radio 4
| | TNC |
| ---------------
|
|
| ---------------
|-----| 'BPQKISS' |
| | |-- Radio 5
| | TNC |
| ---------------
etc
With this system, the Data Engine drives two radio ports (the fastest
or busiest) directly, and has a number of KISS TNCs connected to the
async port. A 3 port node could be built using any KISS TNCs, and a
4 port node would be possible with, for example, a KPC4 (or even another
Data Engine in KISS mode). If you need four or more ports using normal
TNC2's (or clones), you need to run my 'Polled KISS' software, which
allows several TNCs to be driven from one async port, using a simple
diode matrix, and a polling protocol to prevent collisions on the
matrix. The thoretical maximum number of ports is 16, but this might
cause performance problems.
This has the same advantages of reduced NODES size and improved Stats
as the single Data Engine, but also get over the rather nasty problem
of internal routing loops which can occur with multiport NET/ROM
clusters. Although solutions to this problem have been suggested (either
by using a diode maxtrix which does not fully interconnect all the TNCs,
or by careful manipulations of qualities), none seem to be totally
successful. The main disadvantage is that failure of the Data Engine
will disable the entire node complex.
Linked to a PC, which is also running BPQ.
----------------- ---------------
| | | PC |
| Data Engine | KISS |---------------|
| |------------------| G8BPQ | BBS |
----------------- ---------------
| | |
| | |
Radio 1 Radio 2 ? Radio 3 (or more)
The DE async link is defined as a KISS (or NETROM) link, as is one of
the async ports on the PC.
This might make sense if you want to run a node and a BBS, where the node
is carrying so much non-BBS traffic that the PC would not be able to cope
on its own - maybe if the DE was supporting high speed trunk links. The
disadvantages are that the BBS (and SYSOP) cant see the monitored stuff
from the DE radio ports, and that you may end up with more nodes tables
entries than you really need - especially if the PC has other radio ports
attached, and is also running as a switch as well as a BBS. It would
normally make more sense to run the Data Engine in KISS mode (using the
Kantronics EPROM), and let the PC do all the hard work.
John Wiseman, G8BPQ
13/3/91
Updated 5/1/92 and 19/7/92