dias@iris.ucdavis.edu (Gihan Dias) (11/08/90)
I've been following the discussion on MNP, etc. and have a question
about MNP4's operation. I know it packetizes data into "HLDC-like"
frames, and uses an ARQ protocol. What I'd like to know are:
1. What is the maximum frame size?
2. Does it do windowing or just stop-and-wait and, if so, what is the
max window size?
The bottom line is: On an international link which may be subject to
satellite delays, is it better to use MNP4, or turn it off and use a
software protocol like windowed kermit?
Would V.42 be better than MNP4 in this regard?
Thanks for your help.
Gihantnixon@hayes.uucp (Toby Nixon) (11/08/90)
In article <7945@ucdavis.ucdavis.edu>, dias@iris.ucdavis.edu (Gihan Dias) writes: > I've been following the discussion on MNP, etc. and have a question > about MNP4's operation. I know it packetizes data into "HLDC-like" > frames, and uses an ARQ protocol. What I'd like to know are: > > 1. What is the maximum frame size? MNP4's maximum frame size is 256 bytes (of user data), compared to 64 bytes for MNP3. V.42 LAPM's maximum frame size is 4096 bytes, although the default is 128 bytes. > 2. Does it do windowing or just stop-and-wait and, if so, what is the > max window size? Both MNP4 and LAPM are windowing protocols. The maximum window sizes are 127 in both cases, although the "actual" maximum is usually much lower and based on the available memory in the particular implementation (the "maximum" in Hayes V-series modems is 15 frames); the default maximum is 7 frames. > The bottom line is: On an international link which may be subject to > satellite delays, is it better to use MNP4, or turn it off and use a > software protocol like windowed kermit? You'll usually be better to let the modems do the error correction, because the protocol (be it MNP4 or LAPM) has less overhead, usually smaller frames, and can retransmit without the buffering delays associated with a PC-based protocol. A modem-based protocol, properly implemented, can therefore recover faster from line errors than a PC-based protocol. > Would V.42 be better than MNP4 in this regard? Yes, for a couple of reasons. First, it uses poll/final procedures and different frame types for acknowledgement and reject, making it less likely that the intentions of the other modem will be confused (MNP4 uses one frame type for both positive and negative acknowledgement). Second, LAPM, unlike MNP4, has the ability to do Selective Reject, so you can have a large window size and not have to go back a retransmit all those frames if one is damaged in transit. -- Toby Nixon, Principal Engineer | Voice +1-404-449-8791 Telex 151243420 Hayes Microcomputer Products Inc. | Fax +1-404-447-0178 CIS 70271,404 P.O. Box 105203 | UUCP uunet!hayes!tnixon AT&T !tnixon Atlanta, Georgia 30348 USA | Internet hayes!tnixon@uunet.uu.net
root@zswamp.fidonet.org (Geoffrey Welsh) (11/08/90)
Gihan Dias (dias@iris.ucdavis.edu ) wrote: >I've been following the discussion on MNP, etc. and have a >question >about MNP4's operation. I know it packetizes data into >"HLDC-like" >frames, and uses an ARQ protocol. What I'd like to know are: > 1. What is the maximum frame size? > 2. Does it do windowing or just stop-and-wait and, if so, >what is the max window size? I can't answer that directly, but I think that I can correct some impressions that you have (and may Toby correct my every error! <grin>) I do not think that MNP4 in any way regulates the window size. MNP4 is like an 'option' on the base MNP protocols, exactly *one* of which is used as the foundation for any MNP connection. MNP1 through 3, as I understand them, are no windowing, windowing, and windowing with start/stop bit stripping respectively. MNP4 adds to *any* of these the ability to adjust the packet size dynamically. >The bottom line is: On an international link which may be >subject to >satellite delays, is it better to use MNP4, or turn it off >and use a >software protocol like windowed kermit? I have read that throughput may be improved on error-prone links by disabling MNP4, but I have not conducted enough experimentation in this area to provide any observations of my own. >Would V.42 be better than MNP4 in this regard? Again, I can only guess that LAP-M (remember, V.42 *includes* MNP4 in Annex A!) would provide better throughput because it was designed sufficiently later than MNP such that MNP's shortcomings would have been taken into account. Again, Toby would be better qualified to provide a direct answer. -- UUCP: watmath!xenitec!zswamp!root | 602-66 Mooregate Crescent Internet: root@zswamp.fidonet.org | Kitchener, Ontario FidoNet: SYSOP, 1:221/171 | N2M 5E6 CANADA Data: (519) 742-8939 | (519) 741-9553 MC Hammer, n. Device used to ensure firm seating of MicroChannel boards Try our new Bud 'C' compiler... it specializes in 'case' statements!