dplatt@coherent.com (Dave Platt) (03/16/89)
[Please pardon the cross-posting, and note the Followup-to header.]
According to the "Mac the Knife" column in the 3/14 issue of MacWeek,
there's a design peculiarity in the Liason network bridge software that
can have severe repercussions if you use this software with a Telebit
TrailBlazer Plus.
It seems that the software periodically sends a modem-configuration
command string to the modem. According to the article, this occurs once
per minute [I assume that it occurs only when the modem is not "talking
to" another modem... i.e., the bridge is not "alive"]. This
configuration command includes an "&W" sequence, which instructs the
modem to write the configuration information into its nonvolatile
memory.
The TrailBlazer Plus's nonvolatile memory is (according to the article)
implemented in EEPROM, which has a life expectancy of 10,000 write
cycles. At 1 write per minute, Liason could run through this lifecycle
in less than a week! The results would be bad... you might find
yourself unable to change the configuration of your TrailBlazer, or find
that the modem had gone entirely belly-up due to inconsistent EEPROM
settings.
So... if you have Liason and a TrailBlazer Plus, be careful!! You might
want to call InfoSphere and ask for further information and/or a fixed
version of Liason.
Disclaimer: I'm just passing on what I read; I can't attest to the
truth of this report.
--
Dave Platt FIDONET: Dave Platt on 1:204/444 VOICE: (415) 493-8805
UUCP: ...!{ames,sun,uunet}!coherent!dplatt DOMAIN: dplatt@coherent.com
INTERNET: coherent!dplatt@ames.arpa, ...@sun.com, ...@uunet.uu.net
USNAIL: Coherent Thought Inc. 3350 West Bayshore #205 Palo Alto CA 94303alexis@ccnysci.UUCP (Alexis Rosen) (03/19/89)
In article <22320@coherent.com> dplatt@coherent.com (Dave Platt) writes: >According to the "Mac the Knife" column in the 3/14 issue of MacWeek, >there's a design peculiarity in the Liason network bridge software that >can have severe repercussions if you use this software with a Telebit >TrailBlazer Plus. > >It seems that the software periodically sends a modem-configuration >command string to the modem. According to the article, this occurs once >per minute [etc.] This >configuration command includes an "&W" sequence, which instructs the >modem to write the configuration information into its nonvolatile >memory. > >The TrailBlazer Plus's nonvolatile memory is (according to the article) >implemented in EEPROM, which has a life expectancy of 10,000 write >cycles. At 1 write per minute, Liason could run through this lifecycle >in less than a week! The results would be bad... you might find >yourself unable to change the configuration of your TrailBlazer, or find >that the modem had gone entirely belly-up due to inconsistent EEPROM >settings. THIS RUMOR IS ENTIRELY FALSE. I don't know how "Mac the knife" heard this, or why MacWeek was irresponsible enough to print this without verifying it. Liason is an extraordinary product, and this rumor may cause great damage to a company which is doing wonderful and unique things for Mac networking. To Summarize: your Telebits are not at risk. The detail: That command string is necessary to make the modem remember to autobaud. Telebits are shipped so that they default to high-speed mode, which is not what you generally want for a bridge. The write-to-prom command is given ONLY on reset, which generally happens when you reboot your mac. If you reboot five times a day (my gateway, with a telebit, gets rebooted maybe once a week) you won't have to start worrying for about seven years. Long before then Telebit will change their setup. (But before that, you'll likely be on ISDN or some other faster technology [Telebit T5000?].) If any of you remember the horror of dealing with MacServe, you'll know how bad networking can get. As a user of both Liason and (formerly) MacServe, I feel that InfoSphere has more than redeemed itself for that abortion of a disk server. Liason is every bit as good as MacServe was bad. And with version 2.0, which is now in Beta, is will be even better. Alexis Rosen alexis@ccnysci.uucp
dudevoir@isl.Stanford.EDU (Glen P. Dudevoir) (03/20/89)
In article <1422@ccnysci.UUCP> alexis@ccnysci.UUCP (Alexis Rosen) writes: >In article <22320@coherent.com> dplatt@coherent.com (Dave Platt) writes: >>According to the "Mac the Knife" column in the 3/14 issue of MacWeek, >>there's a design peculiarity in the Liason network bridge software that >>can have severe repercussions if you use this software with a Telebit >>TrailBlazer Plus. >> >>It seems that the software periodically sends a modem-configuration >>command string to the modem. According to the article, this occurs once >>per minute [etc.] This >>configuration command includes an "&W" sequence, which instructs the >>modem to write the configuration information into its nonvolatile >>memory. >> >>The TrailBlazer Plus's nonvolatile memory is (according to the article) >>implemented in EEPROM, which has a life expectancy of 10,000 write >>cycles. At 1 write per minute, Liason could run through this lifecycle >>in less than a week! The results would be bad... you might find >>yourself unable to change the configuration of your TrailBlazer, or find >>that the modem had gone entirely belly-up due to inconsistent EEPROM >>settings. > >THIS RUMOR IS ENTIRELY FALSE. > >I don't know how "Mac the knife" heard this, or why MacWeek was irresponsible >enough to print this without verifying it. Liason is an extraordinary >product, and this rumor may cause great damage to a company which is doing >wonderful and unique things for Mac networking. > >To Summarize: your Telebits are not at risk. > >The detail: That command string is necessary to make the modem remember to >autobaud. Telebits are shipped so that they default to high-speed mode, which >is not what you generally want for a bridge. The write-to-prom command is >given ONLY on reset, which generally happens when you reboot your mac. If > >Alexis Rosen >alexis@ccnysci.uucp You haven't explained why the modems are not at risk. If the &w command is given once a minute to insure that the modem autobauds then the result predicted in the column is correct. I should also point out that the same is true for Hayes modems. Either the software gives the &w command once a minute or it doesn't. Which is it? I would like to know since we are considering the purchase of this product for use with hayes modems. Glen
amanda@lts.UUCP (Amanda Walker) (03/20/89)
In article <411@isl.stanford.edu>, dudevoir@isl.stanford.edu (Glen P. Dudevoir) writes: In article <1422@ccnysci.UUCP> alexis@ccnysci.UUCP (Alexis Rosen) writes: >[...] >The write-to-prom command is >given ONLY on reset, which generally happens when you reboot your mac. > >Alexis Rosen >alexis@ccnysci.uucp [...] Either the software gives the &w command once a minute or it doesn't. Which is it? I would like to know since we are considering the purchase of this product for use with hayes modems. Glen And here I thought people read articles that they quoted. Guess I was wrong. Also, the "&W" command is not part of the "standard" AT-command set. Vanilla Hayes modems, in fact, don't even have EEPROMs. You have to configure them the old fashioned way: DIP switches... I still don't particularly see why Liason has to reconfigure the modem on every boot, unless you are reconfiguring your modem a lot when Liason isn't looking :-). It doesn't seem like a problem, though. -- Amanda Walker, InterCon Systems Corporation amanda@lts.UUCP / ...!uunet!lts!amanda / 703.435.8170 -- C combines the flexibility of assembler with the power of assembler.
w-colinp@microsoft.UUCP (Colin Plumb) (03/21/89)
dudevoir@isl.stanford.edu (Glen P. Dudevoir) wrote: > You haven't explained why the modems are not at risk. If the &w > command is given once a minute to insure that the modem autobauds then > the result predicted in the column is correct. I should also point > out that the same is true for Hayes modems. Either the software gives > the &w command once a minute or it doesn't. Which is it? It doesn't. As he said, the &w command is given once per reboot. (I believe Hayes uses battery backup rather than EEPROM, so infinite rewrites are possible.) -- -Colin (uunet!microsoft!w-colinp) "Don't listen to me. I never do." - The Doctor
gdelong@cvman.UUCP (Gary Delong) (03/23/89)
In article <1422@ccnysci.UUCP>, alexis@ccnysci.UUCP (Alexis Rosen) writes: > In article <22320@coherent.com> dplatt@coherent.com (Dave Platt) writes: . > . >It seems that the software periodically sends a modem-configuration . >command string to the modem. According to the article, this occurs once . >per minute [etc.] This . >configuration command includes an "&W" sequence, which instructs the . >modem to write the configuration information into its nonvolatile . >memory. . > > > THIS RUMOR IS ENTIRELY FALSE. > > To Summarize: your Telebits are not at risk. > > The detail: That command string is necessary to make the modem remember to > autobaud. Telebits are shipped so that they default to high-speed mode, which > is not what you generally want for a bridge. The write-to-prom command is > given ONLY on reset, which generally happens when you reboot your mac. If > you reboot five times a day (my gateway, with a telebit, gets rebooted maybe > once a week) you won't have to start worrying for about seven years. Long > before then Telebit will change their setup. (But before that, you'll likely > be on ISDN or some other faster technology [Telebit T5000?].) > > Alexis Rosen > alexis@ccnysci.uucp Well, I wouldn't want any s/w product running on any of my systems that thought it was ok to rewrite firmware. It should be sufficient to inform s/w users that such and such parameters are required for the s/w product to function, and maybe additionally some other parameters would make for most efficient operation. I don't like the idea that some s/w vendor might think that my hardware resources are 'expendable' at his pleasure! I would class a program that, unknown to me, rewrote data in any EEPROM in my system, or in any attached device, as a virus. Does anyone know of any other s/w that issues an AT&W command? -- _____ / \ / Gary A. Delong, N1BIP "I am the NRA." gdelong@cvman.prime.com | \ / COMPUTERVISION Division {sun|linus}!cvbnet!gdelong \____\/ Prime Computer, Inc. (603) 622-1260 x 261
amanda@lts.UUCP (Amanda Walker) (03/24/89)
As several people have kindly pointed out, "AT&W" is in fact a standard command on newer Hayes modems. I hadn't realized this, since the last genuine Hayes I've used was a SmartModem 1200. Sorry about any confusion I may have caused by speaking prematurely. Extracting toes from mouth, -- Amanda Walker, InterCon Systems Corporation amanda@lts.UUCP / ...!uunet!lts!amanda / 703.435.8170 -- "I can only assume this is not the first-class compartment." --HGTtG
TVR%CCRMA-F4@SAIL.STANFORD.EDU (Tovar) (03/26/89)
There certainly has been alot of flaming about Liason software, and when it
appears in print, things definitely get complicated. That's something that
lawsuits are made of. The last message got my attention, as it touches on a
larger issue as well, so here is another flame (and possible solution):
It doesn't _have_ to reconfigure it. In practice, though, they made the right
choice. Two reasons:
[alexis@ccnysci.uucp]
I vehemently disagree that this is the right choice. It definitely should not
automatically reconfigure a modem, and write out that configuration (e.g. alter
the hardware). I speak as a someone who has written terminal simulators and
have dealt with modems and network software before some of the readers of this
list were even in grammar school.
Reaching in and changing system software and especially hardware with explicit
agreement from the user is the sort of thing that viruses are made out of.
There's alot of software out there that does this sort of thing, patching
system calls and the like, and this makes life real hard for someone who has
to maintain the system software, since when there's a bug, it becomes harder
and harder to figure out whose fault it is. An intermittent failure can make
you want to tear your (or someone else's) hair out.
1) They are selling this product to novices as well as experts. Novices will
not know how to configure their Telebit to autobaud, and ... [ibid.]
That's exactly the problem. They make the modem work great with THEIR product
and at the same time, break other products. Novices may not even know what's
wrong, and from what i've seen lately, the average computer store's technician
may have trouble sorting this sort of thing out.
2) Experts may well use the modem with other comm programs at certain times.
Unfortunately, many (if not most) comm programs tell the modem to restore
settings to the factory defaults. So you very well may be "reconfiguring
your modem a lot when Liason isn't looking." [ibid.]
Right again. This restore the modems to some consistent state at which point
the program can configure it for its own purposes. Experts also know what
option work well for their particular brand of modem with the software they
used, neither of which may not necessarily be documented properly. Having this
gratuitously changed is very annoying at best. Thus, it would be nice if, in
general, reconfiguration would be an option. (I'm adding this as an option to
my own program next time i recompile). Much more of this kind of thing out
there, and we'll be seeing `reconfiguation wars'...
My guess is that just because of this flap, they will put in an option to
change this in version 2.0. I'm getting a beta tomorrow, and I'll let y'all
know if it's in the beta. Truth is, I'd just as soon NOT see it. Any "expert"
user paranoid enough to worry can fix it in ten seconds with ResEdit, and it
would clutter up a very well-done interface. [ibid.]
The user should be asked before changing hardware settings, and i hope that
is the option provided. The dialog could simply ask whether to change it:
[Always], [This time], [Never]. Configuring a modem does take time, so
there's reason to avoid it, especially if it needs to happen during the boot
process. Providing these options allows for both speed and compatability.
One shouldn't need ResEdit, a relatively arcane program, in order to do this.
I'm looking forward to seeing Liason correct this defect.
-- Tovar <TVR@SAIL.Stanford.Edu>
P.S. On a related topic, does anyone know of an RS232<->LocalTalk Gateway
which will coexist with an AppleShare server? In the interest of reducing
message traffic, please reply by mail, if possible, and i will post replies
(or summarize) if requested (via mail).jbayer@ispi.UUCP (Jonathan Bayer) (03/27/89)
In article <34@rsoft.UUCP> frank@rsoft.UUCP (Frank I. Reiter) writes: >In article <1083@lts.UUCP> amanda@lts.UUCP (Amanda Walker) writes: >>Also, the "&W" command is not part of the "standard" AT-command set. Vanilla >>Hayes modems, in fact, don't even have EEPROMs. You have to configure them >>the old fashioned way: DIP switches... The Hayes 1200 had DIP switches. The 2400 had EEPROMS. JB -- Jonathan Bayer Beware: The light at the end of the Intelligent Software Products, Inc. tunnel may be an oncoming dragon 19 Virginia Ave. ...uunet!ispi!jbayer Rockville Centre, NY 11570 (516) 766-2867 jbayer@ispi.UUCP
alexis@ccnysci.UUCP (Alexis Rosen) (04/05/89)
I haven't said anything on this subject in a few days because I was doing
some checking.
I have heard from two different people that the Telebit REQUIRES you to
write to the EEPROMs in order to change the speed. This was offered as
justification for their doing this. THEY DO NOT WRITE TO PROMS ON ANY OTHER
MODEM, JUST THE TELEBIT.
Now, that explanation seems VERY odd to me. So sometime in the next few
days, I will fish out a telebit manual and take a look (actually, I am
hoping that a bunch of you will do this and mail to me, but who knows...)
In any event, I want to stress again that whatever they are doing, it is
NOT HARMFUL TO THE TELEBITS!!! They may or may not be doing wrong, but they
WON'T break the modem.
One thing: someone said that he thought programs that wrote to modem proms
were just like viruses. To be equally extreme, I think that that statement
was stupid. The program may be doing something wrong- even extremely
undesireable. But to put that in a class with viruses is ridiculous. If
a newspaper printed something like that, it would likely be legally
actionable.
I will follow up on this within a few days.
---
Alexis Rosen
alexis@ccnysci.{uucp,bitnet}shap@polya.Stanford.EDU (Jonathan S. Shapiro) (04/06/89)
In article <1502@ccnysci.UUCP> alexis@ccnysci.UUCP (Alexis Rosen) writes: >I have heard from two different people that the Telebit REQUIRES you to >write to the EEPROMs in order to change the speed. I have owned a trailblazer plus for some time now, and have set up outgoing uucp to use various different speeds with no trouble at all and no need to rewrite the eeproms. On incoming calls the telebit does automatic baud synchronization. On outgoing calls it also does automatic baud synchronization, and because I lack hardware flow control I explicitly tell it that it should only accept answers at the baud rate I want. One more time: No write to the eeprom is necessary to change the baud rate. Writes to the eeprom *are* required to alter the default answer speed, the answer tone issuance order, and a few default parameters. None of these should be overridden by a program except on a per-call basis. I have configured my telebit to reload the default saved parameters (i.e. the ones I put in the eeprom) whenever DTR or carrier is dropped, and have configured uucp to override the baud rate on a per-call basis. Jon