364 lines
17 KiB
HTML
Executable File
364 lines
17 KiB
HTML
Executable File
<HTML>
|
|
<!-- $Id$ -->
|
|
<HEAD>
|
|
<TITLE>A Type-2 Packet Extension Proposal.</TITLE>
|
|
</HEAD>
|
|
|
|
<!-- Background white, links blue (unvisited), navy (visited), red (active) -->
|
|
<BODY
|
|
BGCOLOR="#FFFFFF"
|
|
TEXT="#000000"
|
|
LINK="#0000FF"
|
|
VLINK="#000080"
|
|
ALINK="#FF0000"
|
|
>
|
|
<PRE>
|
|
Document: FSC-0039
|
|
Version: 04
|
|
Date: 29-Sep-90
|
|
|
|
|
|
A Type-2 Packet Extension Proposal
|
|
Mark A. Howard 1:260/340@FidoNet
|
|
|
|
Status of this document:
|
|
------------------------
|
|
This FSC suggests a proposed protocol for the FidoNet(r) community,
|
|
and requests discussion and suggestions for improvements. Distribution
|
|
of this document is unlimited.
|
|
|
|
Fido and FidoNet are registered marks of Tom Jennings and Fido Software.
|
|
FTS-0001 is a copyrighted work of Randy Bush.
|
|
|
|
Introduction
|
|
------------
|
|
This document serves two major purposes. The first is an attempt to define
|
|
and document the Type-2 packet which is widely in use in FidoNet today.
|
|
Although FTS-0001 defines the structure of a Type-2 packet, the natural
|
|
evolution of our network, mostly with regards to addressing methodology,
|
|
has made it necessary to utilize hitherto unused portions of the header
|
|
to insert Zone and Point information. Also, it has become apparent that
|
|
some of the existing fields are not large enough to accomplish their
|
|
original tasks.
|
|
|
|
The second is to propose a simple mechanism to allow FidoNet to begin to
|
|
utilize advanced mail packing techniques. It is quite apparent that while
|
|
Type-2 has served us faithfully for some time now, the evolution of our
|
|
network in terms of technical and physical complexity has caused us to
|
|
consider more efficient and functional ways to pack mail.
|
|
|
|
It should be made clear that with the exception of the Capability Word,
|
|
Capability Word Validation Copy, ProductCode(hi), and Revision(minor),
|
|
which are proposed extensions to the Type-2 packet header, this FSC is
|
|
an attempt to correctly document existing practices with regards to the
|
|
insertion of zone and point info by at least three mail processors in
|
|
use today.
|
|
|
|
|
|
The Type-2 Header (Where's the Zone?)
|
|
-------------------------------------
|
|
Although FTS-0001 has recently been updated to reflect the use of some of
|
|
the areas in the packed message header for zone data, at least two other
|
|
methods for inserting the zone information have been adopted, making it
|
|
necessary to provide support for both formats in all of the zone aware mail
|
|
processors. The end result is more code, and redundant information in the
|
|
packet header.
|
|
|
|
This has presented a problem in logistics, as it is difficult to consider
|
|
the project of updating mail processors using one type to use the other.
|
|
As sufficient indentification is provided, in the form of the product code,
|
|
to determine the expected location of the zone information, and because
|
|
code already exists in most of the mail processors to overcome this, this
|
|
proposal does not attempt to suggest that one method be used over the
|
|
other, rather the intent is to attempt to document the extensions in use,
|
|
and the products involved.
|
|
|
|
See the accompanying chart and cross-reference.
|
|
|
|
|
|
The Product Code
|
|
----------------
|
|
Based upon the current rate of requests for product codes from the FTSC,
|
|
it is probable that the Product Code byte will not be large enough to
|
|
accomodate all of the codes required. While it is not reasonable to
|
|
expect that all Type-2 processors will eventually check the hi-order byte
|
|
proposed here, it is likely that 'current' mail processors will. This
|
|
can be nothing but benefical, as it will force users to upgrade their
|
|
mail processors to a product which will as a minimum, support Type-2
|
|
with Zone and Point extensions, and quite possibly, processors that will
|
|
utilize more advanced mail packing techniques, making Type-2 extinct once
|
|
and for all.
|
|
|
|
|
|
The Capability Word (How do we GET there from here?)
|
|
-----------------------------------------------------
|
|
Everybody would like to see more efficient and functional ways to pack and
|
|
exchange mail. Several Type-3 message bundle proposals exist, but none
|
|
really address a problem which must be solved first. The problem is that
|
|
since FidoNet is a hobbyist network, no demands can be placed on any one
|
|
sysop to upgrade or change their bundling software. Because of this, it
|
|
is necessary to consider strategies which allow for the existence of Type-2
|
|
bundlers in the network topology.
|
|
|
|
Considerable advantages can be realized, however, between systems that
|
|
consent to use advanced bundling techniques. One way to do this is to
|
|
simply send netmail to all of your connecting systems, saying "Hey, I've
|
|
got a Type-3 bundler now, how about you?" This could become quite
|
|
tiresome, and does not represent much of an improvement on the current
|
|
situation.
|
|
|
|
What would be desirable is a network that would 'upgrade itself'. Given a
|
|
situation where mail processors of various capabilities will exist in a
|
|
network topology, the goal is to provide a mechanism whereby two links can
|
|
determine and utilize the most efficient bundling method to use, in a
|
|
manner transparent to the sysop.
|
|
|
|
For instance, let's say that the FTSC releases the Type-7 All New Singing
|
|
and Dancing bundle format. Well, your current version of SlingToss can
|
|
only support Types 2, 3, and 5. One of your downlinks gets a new version
|
|
of MailMangle which can support Types 2, 3, 4, and 7. Well, it is quite
|
|
obvious that since you and he are exchanging 4 megs of mail each night,
|
|
and it's an overseas call, that it would be in your interest to obtain a
|
|
new version of SlingToss which can support Type 7.
|
|
|
|
Note that this is *optional*. Because both processors can support Type-3,
|
|
they will continue to exchange Type-3 mail quite happily, even though
|
|
MailMangle is happily advertising the availability of Type-7. Even your
|
|
downlinks which are still using stone-age Type-2 processors will be fine,
|
|
as SlingToss will always export Type-2 bundles when no higher capability
|
|
can be determined.
|
|
|
|
So, after dashing off the check to the author, your new version of
|
|
SlingToss comes in the mail! You rush over to your system, and install it.
|
|
The next time SlingToss exports mail to the MailMangle system, it says
|
|
"Hey! I can now support Type 2, 3, 5, and 7! So, whattya got?" This is
|
|
no skin off MailMangle's nose, he's had Type-7 for quite a while, and he
|
|
begins to export Type-7 bundles to SlingToss. "It's about time.", he says.
|
|
|
|
Now, this scenario is made possible by implementing a 'Capability Word' in
|
|
the present and future packet headers. The Capability Update mechanism
|
|
depends on several assumptions:
|
|
|
|
1) Any Advanced Capability Bundler *MUST* be capable of receiving and
|
|
faithfully processing Type-2 bundles. Hopefully, the inbound packets
|
|
will be in the new format proposed by this document, but then again,
|
|
this is not an exact science. What this means is that it is likely
|
|
that some packets may arrive with the Capability Word (CW) set to 0.
|
|
In this case, Type-2 is assumed, assuring compatibility. The only
|
|
caveat is that it is conceivable that some obscure mail processor
|
|
uses the location proposed for the CW for other arcane purposes. This
|
|
| can detected through the CWValidation Copy, which is byte-swapped and
|
|
| compared with the CW at that time. If a mismatch is found, a CW of
|
|
| type 0 is assumed, and a Type-2 bundling method is used.
|
|
|
|
2) An Advanced Capability Bundler, hereafter referred to as a Type-N
|
|
Bundler, must have a method to store and maintain the node-by-node
|
|
capability information. This can be done many ways, and in fact
|
|
several processors already have begun to maintain node information
|
|
outside of that found in AREAS.BBS, mostly to implement pre-arranged
|
|
alternate compression methods. In a text configuration file, you
|
|
might see the following:
|
|
|
|
; Address Comp Send LastCW ; Comments
|
|
Node 1:260/340 ZIP Auto 7 ; Auto detect & upgrade
|
|
Node 1:135/20 LZH 3 2,3,7 ; Always send Type-3
|
|
Node 1: ARC 2 0 ; Stone-Age processor
|
|
Node 1:135/4 --- Auto 7 ; Sent me netmail
|
|
Node 1: --- 0 0 ; Don't send CW
|
|
|
|
In this example, the fields are:
|
|
|
|
Address - downlink address. Note that this is not necessarily
|
|
relative to echomail, only, it is possible to append
|
|
information to the node database as netmail packets are
|
|
receieved from different addresses.
|
|
|
|
Comp - desired mail compression method.
|
|
|
|
Send - Auto - automatically determined maximum common packing
|
|
method to use. Automatically update to LastCW
|
|
when packing.
|
|
|
|
LastCW - Last CW received from remote system.
|
|
|
|
|
|
3) A Type-N Bundle will always advertise it's capabilities in the CW
|
|
regardless of the type being sent. As shown in the above example,
|
|
it allows Type-N processors to automatically track the capability
|
|
of your system. Again, in cases where a stone-age processor is
|
|
being used, this field will be ignored, and in the unusual event
|
|
that it is not ignored, and is somehow harmful to the far system,
|
|
the Type-N processor can be configured to send a CW of 0.
|
|
|
|
The format of the Capability Word is designed to support up to 15 future
|
|
bundle types, and is bit-mapped to facilitate the easy determination of
|
|
the maximum common level supported between two nodes:
|
|
|
|
msb Capability Word lsb
|
|
Node Supports ------------FTSC Type Supported----------------
|
|
|
|
U 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2
|
|
|
|
2, 3, and 7 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 1
|
|
2, 3, and 5 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 1
|
|
2 (this FSC) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
|
|
Stone Age** 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
|
|
^
|
|
+--- Indicates UseNet RFC-822 capability
|
|
|
|
** - a mismatch in the CWValidation Copy also
|
|
produces a CW=0.
|
|
|
|
In this example, the Type-N bundler would first compare the remote CW
|
|
| and the byte-swapped remote CWValidation Copy, and check for a mismatch.
|
|
Prior to the compare, the MSB of the CW's are masked, as this bit is
|
|
reserved to indicate whether the mail processor is capable of also
|
|
accepting UseNet RFC-822 bundles. Following the MSB mask, and bit
|
|
comparison, if they do not match, a remote CW of 0 is assumed. If they
|
|
match, the Type-N processor would AND the local and remote CW words,
|
|
obtaining a CW expressing the Types which are in common to both systems.
|
|
The most significant Type will be used, by default, but note that this
|
|
assumes that new bundling Types will be increasingly more efficient or
|
|
in some way more beneficial.
|
|
|
|
Because this may not always be the case, there should be a method provided,
|
|
as illustrated above, to override the automatic upgrade should this become
|
|
the case.
|
|
|
|
The MSB of the CW is used to indicate whether the mail processor can accept
|
|
UseNet RFC-822 bundles or not. It is a separate indicator, and not intended
|
|
to be used as a part of the above comparison, however can be used to also
|
|
advertise RFC-822 capability to other systems. Since RFC-822 is 'set in
|
|
stone', there is no need to assign more than one capability bit.
|
|
|
|
It might seem somewhat limiting to only consider the possibility of 15
|
|
different future bundling methods, but it is my opinion that the careful
|
|
use of a 'Sub-Type' byte in the packet header can allow for the variations
|
|
on a single theme, and that proposals for new formats should be evaluated
|
|
by the FTSC to determine whether sufficient benefit can be realized in
|
|
the implementation of the given format, prior to assigning a new type
|
|
code.
|
|
|
|
|
|
Mailers
|
|
-------
|
|
It is quite clear to me that should this concept take hold, that the
|
|
logical place to store node capability data is in the local nodelist
|
|
database, or an index-associated data file. As above, it is necessary
|
|
to generate Type-2 packets for whatever purpose, unless it is known
|
|
by prior association, that the far mailer can accept other types of
|
|
packets. It is also quite reasonable to assume that a nodelist flag
|
|
could be assigned to advertise the CW for a given node, which the
|
|
native mailer nodelist compiler could then immediately determine the
|
|
preferred bundling method for any given node in FidoNet.
|
|
|
|
Another possibility would be to pass a capability advertisement in the
|
|
extensible portion of a handshake protocol, which may or may not already
|
|
exist in FidoNet.
|
|
|
|
The approach suggested previously in this document suggests the use of
|
|
a text configuration file, but it is quite obvious that many benefits
|
|
can be realized through the use of the nodelist, including the use of
|
|
additional flags to indicate the preferred compression method, etc.
|
|
|
|
|
|
Summary
|
|
-------
|
|
This document has been created in an attempt to define a method to allow
|
|
the future expansion and enhancement of FidoNet technology mail processors
|
|
and mailers, and in a way that is the least disruptive to existing mail
|
|
operations. The intent is to provide for an environment that is as open,
|
|
and extensible as possible.
|
|
|
|
The mechanism described should allow many different types of processors
|
|
(FTSC-registered) to exist in the network at once, and to provide an
|
|
environment which is designed to operate at it's maximum efficiency
|
|
wherever possible or practical.
|
|
|
|
Revision 2 of this document was produced to implement suggestions made
|
|
primarily by Jan Vroonhof, who suggested the use of the CW Validation
|
|
Copy. Jan presented this idea in his FSC-0048, along with other concepts
|
|
relating to the correct indentification and handling of zone and point
|
|
addressing. This document sanctions the improvements to the CW as
|
|
recommended, but does not address or support the other extensions
|
|
recommended in FSC-0048.
|
|
|
|
|
|
Thanks
|
|
------
|
|
To Ward Christensen, creator of XModem and BYE.
|
|
|
|
Tom Jennings, who started this whole mess.
|
|
|
|
Joaquim Homrighausen, for lots of good ideas, and motivation. Here's
|
|
another Lamborghini to work on.
|
|
|
|
Wynn Wagner, Oliver McDonald, Roeland Meyer, Andrew Farmer, Claude
|
|
Warren, Jan Vroonhof, Bob Hartman, and Vince Perriello, who all
|
|
contributed in some way to the creation of this document, mostly
|
|
through their messages in NET_DEV.
|
|
|
|
|
|
|
|
Type-2 Packet Format (proposed, but currently in use)
|
|
-----------------------------------------------------
|
|
Field Ofs Siz Type Description Expected value(s)
|
|
------- --- --- ---- -------------------------- -----------------
|
|
OrgNode 0x0 2 Word Origination node address 0-65535
|
|
DstNode 2 2 Word Destination node address 1-65535
|
|
Year 4 2 Int Year packet generated 19??-2???
|
|
Month 6 2 Int Month " " 0-11 (0=Jan)
|
|
Day 8 2 Int Day " " 1-31
|
|
Hour A 2 Int Hour " " 0-23
|
|
Min C 2 Int Minute " " 0-59
|
|
Sec E 2 Int Second " " 0-59
|
|
Baud 10 2 Int Baud Rate (not in use) ????
|
|
PktVer 12 2 Int Packet Version Always 2
|
|
OrgNet 14 2 Word Origination net address 1-65535
|
|
DstNet 16 2 Word Destination net address 1-65535
|
|
PrdCodL 18 1 Byte FTSC Product Code (lo) 1-255
|
|
* PVMajor 19 1 Byte FTSC Product Rev (major) 1-255
|
|
Password 1A 8 Char Packet password A-Z,0-9
|
|
* QOrgZone 22 2 Int Orig Zone (ZMailQ,QMail) 1-65535
|
|
* QDstZone 24 2 Int Dest Zone (ZMailQ,QMail) 1-65535
|
|
Filler 26 2 Byte Spare Change ?
|
|
| * CapValid 28 2 Word CW Byte-Swapped Valid Copy BitField
|
|
* PrdCodH 2A 1 Byte FTSC Product Code (hi) 1-255
|
|
* PVMinor 2B 1 Byte FTSC Product Rev (minor) 1-255
|
|
* CapWord 2C 2 Word Capability Word BitField
|
|
* OrigZone 2E 2 Int Origination Zone 1-65535
|
|
* DestZone 30 2 Int Destination Zone 1-65535
|
|
* OrigPoint 32 2 Int Origination Point 1-65535
|
|
* DestPoint 34 2 Int Destination Point 1-65535
|
|
* ProdData 36 4 Long Product-specific data Whatever
|
|
PktTerm 3A 2 Word Packet terminator 0000
|
|
|
|
* - extensions to FTS-0001
|
|
|
|
Ofs, Siz are in hex, other values are decimal.
|
|
|
|
|
|
Zone/Point Aware Mail Processors (probably a partial list)
|
|
----------------------------------------------------------
|
|
Prod
|
|
Code Name - Uses QOrg/QDstZone Orig/DestZone Orig/DestPoint
|
|
---- ----------- ------------- ------------- --------------
|
|
0x0C FrontDoor Reads/Updates Yes Yes
|
|
0x1A DBridge ????? Yes Yes
|
|
0x45 XRS Reads/Updates Yes Yes
|
|
0x29 QMail Yes ????? Not point-aware
|
|
0x35 ZMailQ Yes ????? Not point-aware
|
|
0x3F TosScan Reads/Updates Yes Yes
|
|
|
|
|
|
|
|
|
|
|
|
</PRE>
|
|
|
|
<A HREF="index.htm"><IMG SRC="../images/b_arrow.png" ALT="Back" Border="0">Go Back</A>
|
|
|
|
</BODY>
|
|
</HTML>
|
|
|