| |
SM7UZI > BPQ 29.01.99 13:26l 325 Lines 13940 Bytes #-10126 (0) @ WW
BID : SM7UZI4778
Read: SP6VXT DB1RAS GUEST
Subj: Re: bpq vs flexnet
Path: DB0AAB<DB0PV<DB0WGS<OE5XBL<DB0RGB<DB0ABH<DB0SRS<DB0MW<DB0ERF<DB0HSK<
PI8DRS<DB0PKE<DB0SM<DB0NDR<DB0HES<OZ7BOX<OZ5BBS<OZ4PAC<OZ4BBS<OZ4BOX<
OZ3BOS<OZ3PAC<OZ6BBS<OZ3BOC<SK7IQ
Sent: 990128/1519Z @:SK7IQ.MALMOE.M.SWE.EU #:358 [Malmö] $:SM7UZI4778 FBB7.00g
From: SM7UZI@SK7IQ.MALMOE.M.SWE.EU
To : BPQ@WW
Hello.
>>Be happy about that you have the opportunity to get information that you
>>don't have the ability to collect by your self.
>Please keep quiet then.
The bulletins are not for only for you.
>Why tell us anything at all if it isn't your duty to tell us anything ?
>All the information you has given us is "abandon BPQ, FlexNet is better".
No not all the information, but one conclusion.
>Real information, with facts, is collectable from the developers
>web-pages. (And I have the opportunity to collect the information that way)
The do it, and even learn something maybe.
Also from other webpages are good information avaliable.
>Guess again.....
>I'm not conservative, and I have actually much knowledge what packet is,
>and what can be done do improve it.
>I have both technichal knowledge, practical knowledge, experimental knowledge
>and theoretical knowledge. I think I know what packet is....
You write very many strange things, that makes me belive that you don't realy
understand.
>I have never said bpq is better than another node-software.
It is less good than most others.
>Just replying to FlexNet-fans to correct them then they are wrong about
>something regarding BPQ or KISS, and asking questions to clarify their
>statements.
You are replaying even if they are right adn not directly clearifuing any
thing.
>KISS is still no protocol...
The protocol used for communication between the tnc and host when using kiss is
the protocol.
>I don't think 6PACK is any protocol either, it is a host-mode that user
>AX25 as a protocol.
No is isn't.
Its function is the same as the function of kiss.
>You can use high-speed-TNC's with BPQ, just connect it to the computer.
I mean in a way that will work good!
>This is a setup-problem in BPQ, and other nodes.
>You can tell the node to use another route instead of the one choosen by
>the node.
That is something you don't have to do when uing flexnet.
To let flexnet handle the routing is much more efficient then any manual setup.
>Network-updates can be done as often you like.
The typpical value mentioned in bpsq docs is 60m, that is what is used
often. The whole network must update the netwrk inforation often if it
shall make any difference.
>You are clearly talking about a network there a node has a LOT of
>neighbours. In this situation I can see the advantage, but not in a system
>where it is a significant distance between nodes, and you have a
>"high-way" to transport packets over. (This is the reality in many areas.)
It would be good also then.
Better than using bpq.
But bpq still doesn't work even if there are very few neighbours, more than
two and the routign is messed up.
Some times does it route a compleatly wrong path.
>In my experience, a computer-node is not suitable for an unattended node.
No, then is the rmnc a good solution.
>If the computer hangs for some reason, you must restart it manually.
>The choice of TheNet in a TNC2, or clone, is fool-proof.
>You can use what speed you like, 0 bps to 64Kbps, just replace the
>modem in the TNC.
TNC2 isn't capable of handling a good packet driver or node software, one
similair to a node software for computers.
But the rmnc hardawre is, and it is good both for low and high speeds.
>What bullshit am I loading the network with ?
You write many strange thing,the reason is probably a missunderstanding or
lack of knowledge.
>I have never said that I am providing any information.
>Just correcting errors about BPQ and KISS, and asking questions.
What are the errors then?
If you have used bpq and kiss must you have experianced many of the problems
mentioned.
>I have read ALL bulletins in this subject.
>It is mainly you on the FlexNet-side, and a lot of others that are making
>various kind of statements and asking questions.
That is nice.
>>>Please excuse me, but KISS has a standard.
>>Who have standardised it?
>K3MC, Mike Chepponis and KA9Q, Phil Karn invented KISS in 1987.
That doesnt make it a standard.
There are many variants like smack, bpqkiss, rmnc-crc32 kiss and also others.
Not realy a standard.
>KISS works this way:
>You create the AX25-frame in the application, with headers and all.
>This makes it possible to use other protocols than AX25 if you like.
If you know how it works and have experiance of using it how come that you
don't realize what the problems are and why 6pack is to prefer?
>It is NOT, and has NEVER been a protocol.
>It is old, but not worthless.
It is worthless for use with flexnet, sicne flexnet would work less good.
And there are better ways of using a tnc, like with 6pack, in or in hostmode.
>It is a modem with a HDLC-driver.
>It makes a cheap TNC emulate any expensive TNC you like, as the protocol
>is in the computer-software.
That works less good than a much cheaper modem and requires a comptuer to work.
>Why use an expensive modem when you can do it all with a cheaper one ?
Yes why?
The only thing in the tnc used is the modem!
>>>To really use the standard is perhaps not that easy, as you must fully
>>>understand the packet-protocol you want to use.
>>Keep it simple stupid!
>You said it. Why use more complex things than you have to ?
Because it is to simple to work good.
>Why is it used then ?
Because many tncs support it and it has been around for a long time.
Their tncs aren't good enough to handle a packet driver or node with the
capabillities of a comptuer based.
And pople often doens't know what the disadvantages of using it are.
Most packet users realy don't know any thing about kiss, tncs packet drivers
or any thing else about packet radio.
>Original KISS sends the packet-frame to the application as soon it recieved
>a valid frame.
It often not works the way it is supposed to.
>Host-modes is no protocols, it uses TNC's with pre-installed AX25-software.
>HDLC is no protocol, it is a techique to modulate the packet-frame.
Yes.
>I fail to see how the comparison can be made.
>The comparison is between a hostmode-protocol, and a method to let the
>packet-protocol reside in software.
6pack is no hostmode.
Flexnet is a packet driver and it is either handled by a computer or a rmnc
hardware.
>All the drawbacks that KISS is suffering of, according to the person doing
>the comparison, is easily corrected if you program the application
>the correct way.
How should that be done!
One problem is that the packet driver isn't aware of the dcd, ptt status etc.
Not without using something else than kiss, like 6pack any way.
>If you make an application that emulates 6PACK in software, you kan use
>KISS as an interface to the radio. The behavior will be EXACLY as is the
>original 6PACK.
That is a very stupid solution that wouldn't work either!
>As you can't compare the two things, I can't answer that.
>But I hope I have managed to tell you what KISS is, and what you can do
>with it.
You write rather much about things that you obviosly don't know very much
about.
But how you can avoid to understand the problems with kiss and bpq if you have
read all the bulletins doesn't I under stand.
>>Things have changed since than, other better ways to use a tnc have been
>>discovered.
>Like using it as a brickin a brick-wall ?
>What else can you do with a TNC beside as a link between the user and the
>radio ?
There have been new apcket drivers devloped that requires another way of
controling the tnc and the modem in it.
You can't use a tnc2 for as standalone nodes any more.
Not in host mode, since the packet drivers that the tnc is capable of handle
is not as good as the packet drivers handled by computers any more.
Even if most people doesn't use any of the new good packet drivers yet.
>>>With KISS you have transferred the packet-protocol to the computer-software.
>>But not the full control of the modem used!
>Yes you have, you can attach whatever modem you like to a KISS-TNC, or
>SCC-card using a KISS-driver.
>If you have more than one modem, you have one KISS-port for each of them.
>(BPQ-KISS can use multiple KISS-TNC's/port)
That doesn't solve the problem!
The problem is that the packet driver doesn't control the tnc, and that the
tnc doesn't provide the packet driver with real time inforamtion.
>They don't have to think about the packet-protocol.
>They are using a simple solution, letting others take the blame if something
>doesn't work in the transport-layer.
>"The driver doesn't do what my application tells it to do..."
The packet driver handles the communication, every one must not make their own
packet driver becsue they write a packet radio software.
To use a already excisting packet driver is often the best sollution.
To write other kiss using software wouldn't solve that.
>You doesn't have to control a modem as it just is a modulator/demodulator.
>The software is controlling HOW it is used.
The problem is that if you use kiss can't the software in the computer do that.
>I use 4.08a, and I have never notised this "bug".
>What operatingsystem was used ? (DOS, Windows, OS2 ?)
ms-dos.
The routing problem and kiss problesm must you ahave noticed.
If you know anything about ax25 and the networks you are using.
>Was BPQ properly configuerd to use 2 calls ?
>Not using the same ALIAS on both ports ?
The node does only have one alaias.
>You are makeing a Level2-connect that way.
>If you must do that to reach a node directly, that node must be contantly
>going up and down, messing up the routing-lists.
No.
Bpq missroutes quiet alot.
>This is corrected by locking the route to a specific port, and neighbour.
That really doesn't help when a node far away connects its neighbour and the
frames are routed a compleately wrong way.
The whole point in using net/rom it that the nodes shall handle the routing.
A net/rom node shall determine the best way to route.
That doesn't work, but the problem is solved by using another node like
flexnet.
>>when calling for a station can you use whatever name you want.
>>But when the stations are connected must the callsigns be used,
>>so it must answer with a callsign.
>
>Please tell me where I can find that change in the regulations.
>If you are right, I'm sorry...
You are wrong.
>If you, as a user, connects to a node using its ALIAS, it replies with the
>ALIAS, not the call-sign. The nodes call-sign is never in any frame in
>that session.
That is right and a problem of net/rom nodes.
But to use the alias is may times the only way to connect, since bpq are used.
>I fail to see what you mean with that reply.
>I don't see what BPQ has to do with that it must be a real call in the
>address-field.
When using the alias for connection does bpq answear, not with its call sign
but as if the alias was the callsign.
During the communication shall callsigns be used, but aliases are used
instead..
>>> My recepy is: Never use alias in a connect-sequence.
>>That doesn't help much!
>
>Yes, if the configuration is correct.
To use port calls can help.
But they are often only used to connect the bbs.
To be able to get around the poor routing must often aliases be used.
>This indicates a few things:
>1. Too many nodes in the same area, on the same frequency.
No, that is not the problem.
To have many nodes should be good, since it will increase the capabillity of
the network.
>2. Badly configured nodes.
Not in this case i think.
But that is a big problem often.
A problem you don't get when using flexnet.
>3. Nodes that keeps going up and down, messing up the routing.
No but the link quallity is not constant.
The meaning of bpqs quallity measurement is that is always shall route the
best way but it doesn't.
> A node just have to be up for a second, and the routing is there for
> a long time. Nodes shouldn't go up and down.
That is a problem that comes with bpq.
If flexnet had been used had a lost node been out of the node list in a few
minutes.
>If a node has a poor path, and you get the same phenomena as i reason 3,
>you should lock that route as "no-route".
The nodes shall be able to rotue the best way.
When there are much traffic on yor locked path is it better to use another
path with less trafic, but you stop the node from doing that.
To let the node handle it is the whole point, if that doesn't work are there no
meaning in the use of a net rom/node.
The solution is to use another node.
>Routing-problem can be solved by proper configurations.
Not the problem that bpq can't handle the routing properly.
We are then back to the old days of tcp/ip when no automatic routing was used.
You can't ever confugerue bpq in such way that it works as good as flexnet.
>TheNet, the TNC-version, and BPQ has never been PLUG-N-PRAY.
No and now are they not as good as other avaliable nodes any more either.
>You have to configure it manually to make it work optimal.
But you can never get it as optimally working as flexnet and its adaptive
parameters.
>A lot of node-sysops doesn't do that, ant you doesn't get an optimal network.
Most doesn't, and every one will suffer.
This is a big problem and a node that is easy to setup will help.
>Badly conigured nodes is not any reason to violate any regulation or law.
The problem is bpq.
To use bpq in a way that works are then often agaisnt the law.
.a&$a. .a&$a. .a&$$$$$$$$$$&a. .a&$a. Anders Lageras
$$$$$$ $$$$$$ $$$$$$$' `$$$$$$ $$$$$$ Lund Sweden JO65PQ
$' `$$ $$$$$$ `ý$$$$$s,s$$$$$ý $$$$$$ AARS Member No. 81
$s,s$$ $$$$$$ a$$$$$$ý $$' `$ Packet:
73 de SM7-$$$$$$ $$$$$$ a$$$$$$ý $$s,s$ SM7UZI@SK7IQ.MALMOE.M.SWE.EU
$$$$$$ $$' `$ a$$$$$$ý $$$$$$ E-mail: anders.lageras@usa.net
$$$$$$a$$s,s$ a$$$' `$$$$$$&a. $$$$$$ ICQ#: 5161379
$$$$$$$$$$$$$ $$$$s,s$$$$$$$$$ $$$$$$ Homepage:
`ý&$$$$$$$$ý' `ý&$$$$$$$$$$$ý' `ý&$ý' http://www.visitweb.com/uzi/
o==|=====-
Read previous mail | Read next mail
| |
