home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.dcom.modems:11136 can.uucp:171
- Path: sparky!uunet!uunet.ca!xenitec!zswamp!geoff
- From: geoff@zswamp.UUCP (Geoffrey Welsh)
- Newsgroups: comp.dcom.modems,can.uucp
- Subject: Re: UUCP 'g' vs. MNP & V.42
- Message-ID: <mN8FoB2w164w@zswamp.UUCP>
- Date: Fri, 24 Jul 92 23:31:45 EDT
- References: <1992Jul23.180658.20724@eci386.uucp>
- Organization: Izot's Swamp
- Lines: 88
-
- Just the facts so's the folks make the best decisions:
-
- woods@eci386.uucp (Greg A. Woods) writes:
-
- > MNP-4 is link-layer protocol that retries erroneous packets. UUCP 'g'
- > is a link-layer protocol that retries erroneous packets. Guess what
- > -- the don't get along very well!
-
- This is *completely* untrue! While MNP may increase latency compared to a
- normal link, it does not in any way interfere with the data being sent.
- UUCP-g should not have to retry erroneous packets because MNP will correct
- them, but what harm is there in that? In addition, MNP3 (the base protocol
- usually used on MNP links; MNP4 is an additional feature allowing dynamic
- packet sizing) strips start and stop bits, increasing throughput as much as
- 20% over normal asynchronous links, even after the error correction overhead
- is taken into account.
-
- There is no harm in using an error-correcing protocol over an 'error-free'
- link (I put that in quotes because MNP guarantees no errors between the
- modems, not no errors between the computers), unless the protocol imposes a
- significant overhead. Since, in this case, the overhead is more than made up
- for by start and stop bit removal, there is no reason not to use MNP under
- UUCP-g, at least not at 2400 bps where the increased latency should not push
- past the window size.
-
- > MNP-5 is a compression scheme. I don't believe you can have MNP-5
- > without also using MNP-4, i.e. don't try to use MNP-5 with UUCP 'g'.
-
- MNP5, like MNP4, requires that one of MNP1 through MNP3 be active. I am
- not aware of anyone adapting MNP5 to work with any other link layer protocol.
-
- > On the other hand, you do want to *allow* (but not force) MNP for
- > dial-up, in case it's a human, or co-operating protocol like SL/IP.
-
- Why would you think that SLIP would "co-operate" with MNP while UUCP-g
- would not?!? Doesn't SLIP also resend lost/corrupted packets? What, then, is
- the *mechanical* difference between the two?
-
- > BTW, so far as I know, the V.42bis compression protocol is quite
- > useful in conjunction with UUCP 'g', especially on a modem that does
- > local spoofing of the 'g' protocol (such as the T-2500).
-
- But V.42bis is an data compression protocol which, like MNP5 compression,
- must be implemented over a "link-layer" error correction protocol (normally,
- V.42... which, although not compatible with MNP3/4, is functionally
- equivalent). If V.42 error correction is benign to UUCP-g (and I'm certain
- that, at 2400, it's significantly *beneficial*), then MNP4 must also be.
-
- What you may be confusing this with is the problem when MNP5 gets its hands
- on precompressed data. MNP5 is a 'dumb' compression protocol in that it does
- not attempt to assess its effectiveness; when sending precompressed data (e.g.
- compressed newsbatches), it can actually *expand* the data rather than
- compress it. It has therefore been known that, when you are expecting to
- transfer compressed news, you should disable MNP5... but leave MNP4 on.
-
- V.42bis, on the other hand, does in fact keep track of its effectiveness.
- If it detects that it is expanding data, it will actually turns data
- compression off to prevent a negative effect on throughput. It is therefore
- harmless to leave V.42bis data compression on, no matter what kind of data
- you're sending.
-
- Sadly, few modems are configurable to allow V.42bis data compression but
- not MNP5; USRobotics modems are among those few.
-
- > The Addendum for the WorldBlazer talks about the Telebit LZ Data
- > Compression scheme, and hints that V.42bis can be used with PEP. It
- > does not say if Tb's LZ is "safe" with pre-compressed data (like LAP-M
- > is). [V.42bis, MNP-5, and LZ are all variants of the Lempel-Ziv
- > compression algorithm, though most implementations of MNP-5 *reduce*
- > throughput with pre-compressed data, and each have varying degrees of
- > efficiency and ratios of compression.]
-
- Correction: MNP5 is *not* a Lempel-Ziv based algorithm. It is based on the
- Huffman algorithm (a byte frequency based system), while MNP7 performs Huffman
- coding on byte pairs to achieve better compression. As mentioned above,
- V.42bis (AKA British Telecom Lempel-Ziv, or BTLZ) monitors its effectiveness
- and will shut itself off if it detects that it is expanding data. LZ
- compression is not inherently "safe" with precompressed (or random) data.
-
- V.42bis can be adapted to any error correction scheme, as USRobotics
- proved: they implemented it on top of their HST link-layer protocol (which is
- non-standard - I believe an extension of MNP - and includes provision for
- controlling the reversal of the high speed carrier direction).
-
- It's amazing what you can learn from comp.dcom.modems. <grin>
-
- Geoffrey Welsh, 7 Strath Humber Court, Islington, Ontario, M9A 4C8 Canada
- geoff@zswamp.uucp, [xenitec.on.ca|m2xenix.psg.com]!zswamp!geoff (416)258-8467
-