drubin@prism.poly.edu (Dave Rubin) (05/08/91)
I am in the process of upgrading our dialup modems, most likely with some combination of 9600bps and 2400bps modems. Our telephone switch cannot handle async speeds over 19.2kbps. Therefore, the potential 38.4kbps throughput of V.32/V.42bis modems would probably be wasted. Is there any other reason to go with V.42bis on our V.32 modems? As for 2400bps, for standard dialup applications, is there much to gain by choosing V.42bis over MNP5? We plan on setting the DTE rate at a constant 9600bps regardless of the actual speed of the connection. Finally, does anyone have a recommendation for inexpensive (but reliable) V.32 and 2400bps modems, preferably rack-mountable, that can be set to a "manual" mode (where the modem plays dumb and allows our PBX to control it). Any info would be appreciated... -- Dave Rubin Polytechnic University drubin@prism.poly.edu
tnixon@hayes.uucp (05/08/91)
In article <1991May7.183758.14662@prism.poly.edu>, drubin@prism.poly.edu (Dave Rubin) writes: > I am in the process of upgrading our dialup modems, most likely with some > combination of 9600bps and 2400bps modems. Our telephone switch cannot > handle async speeds over 19.2kbps. Therefore, the potential 38.4kbps > throughput of V.32/V.42bis modems would probably be wasted. Is there any > other reason to go with V.42bis on our V.32 modems? Certainly! If you had no data compression at all, you'd only get about 1150cps max. With MNP5 you could approach 1920cps on highly-compressable files, but are more likely to see in the neighborhood of 1500-1600 on "typical" text -- and MNP5 will have a negative impact (less than 1150 cps) on throughput while sending uncompressible data. With V.42bis, you'll see the full 1920cps through much more of the time. Even though the algorithm is capable of much more, it's not "wasted" because it does allow you to take advantage of all the bandwidth you actually have available. Since many companies sell V.42bis modems for no more than they sell MNP5, you might as well get V.42bis. > As for 2400bps, for standard dialup applications, is there much to gain by > choosing V.42bis over MNP5? We plan on setting the DTE rate at a constant > 9600bps regardless of the actual speed of the connection. Absolutely. Again, MNP5 gets about 2-to-1 (480cps) on highly compressable data, but more like about 350-400 on "typical" data. V.42bis will typically do at least 50% better (600-700cps on "typical" data) and 900cps+ on well-compressible data. I can definitely perceive the difference between V.42bis and MNP5 on the same data. > Finally, does anyone have a recommendation for inexpensive (but reliable) > V.32 and 2400bps modems, preferably rack-mountable, that can be set to a > "manual" mode (where the modem plays dumb and allows our PBX to control it). Since I work for a modem manufacturer, I'll let other folks make the recommendations. :-) However, I will point out that FEW of the V.32 multimode modems with V.42bis on the market today support MI/MIC operation (which is what you want); the Hayes Ultra 96 is oen that does. Perhaps some users of other modems can check their spec sheets and see if they support MI/MIC. -- Toby Nixon, Principal Engineer | Voice +1-404-840-9200 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.uucp (Geoffrey Welsh) (05/08/91)
In a letter to All, Dave Rubin (drubin@prism.poly.edu ) wrote: >I am in the process of upgrading our dialup modems, most >likely with some combination of 9600bps and 2400bps modems. >Our telephone switch cannot handle async speeds over 19.2kbps. I'm not sure that this has any effect on the choice of modems, unless the modems will be used to transmit data generated by the telephone switch itself, which is unusual or possible. If the telephone switch is just a PBX and the modem's serial port is being plugged into another computer, then this isn't an issue at all. If the switch is digital and has a low sampling rate, you may have a problem with higher speed modems; only experimentation (or the voice of experience) will be able to tell what the maximum reliable speed will be. In any case, the maximum physical bit rate of V.32 modems is 9600, and of V.32bis modems it's 14400. V.42bis doesn't increase the speed at which the two modems communicate with each other, and its presence or absence should have no effect on whether a line will support a certain speed. >Therefore, the potential 38.4kbps throughput of V.32/V.42bis >modems would probably be wasted. Keep in mind that this "potential" is the result of data compression, which does not in any way affect the rate the real data bits are passed through the phone line! -- Geoffrey Welsh - Operator, Izot's Swamp BBS (FidoNet 1:221/171) root@zswamp.uucp or ..uunet!watmath!xenitec!zswamp!root 602-66 Mooregate Crescent, Kitchener, ON, N2M 5E6 Canada (519)741-9553 "He who claims to know everything can't possibly know much" -me
drubin@prism.poly.edu (Dave Rubin) (05/10/91)
>>I am in the process of upgrading our dialup modems, most >>likely with some combination of 9600bps and 2400bps modems. >>Our telephone switch cannot handle async speeds over 19.2kbps. > > I'm not sure that this has any effect on the choice of modems, unless the >modems will be used to transmit data generated by the telephone switch itself, >which is unusual or possible. > > If the telephone switch is just a PBX and the modem's serial port is being >plugged into another computer, then this isn't an issue at all. > > If the switch is digital and has a low sampling rate, you may have a >problem with higher speed modems; only experimentation (or the voice of >experience) will be able to tell what the maximum reliable speed will be. Well, I guess I should elaborate. Our telephone switch (Intecom IBX) is digital, and supports async data on an RS-232 box attached to the telephone. The modems are used in a pool, and are allocated by the switch as needed, only for outside calls. Data transmission within the PBX is entirely digital. Modems are not attached directly to the telephone, in fact, this is impossible since the phones are digital. Therefore it is up to the PBX to transmit data between the modems and the async devices attached to a user's computer, and this is limited to 19.2k. >In any case, the maximum physical bit rate of V.32 modems is 9600, and of >V.32bis modems it's 14400. V.42bis doesn't increase the speed at which the >two modems communicate with each other, and its presence or absence should >have no effect on whether a line will support a certain speed. The speed limitation I refer to is on the DTE side, not the analog side. Since V.32/V.42bis could potentially achieve 4:1 compression, I would assume that to make full use of this potential the DTE would need to be set to 38.4k, which is not possible on our PBX. >>Therefore, the potential 38.4kbps throughput of V.32/V.42bis >>modems would probably be wasted. > Keep in mind that this "potential" is the result of data compression, which >does not in any way affect the rate the real data bits are passed through the >phone line! But it does affect the speed that the DTE port needs to run at, correct? I am pretty new to these new standards, so please let me know if I am missing something. -- Dave Rubin Polytechnic University drubin@prism.poly.edu
root@zswamp.uucp (Geoffrey Welsh) (05/11/91)
Now that you've described your needs in greater detail, I think I can provide some useful input. Even if your async speed is limited to 19.2kbps, it may pay you to invest in the fastest modems you can get a hold of. Data comression is a big advertising feature these days, but its effectiveness varies from the claimed 4:1 potential right down to zilch... a V.32bis modem, operating at 14.4kbps physical port speed, won't exceed a throughput of 1800 CPS on compressed data, with or without V.42bis data compression. If you *need* the maximum throughput possible, buy the fastest dialup modems out there (v.32bis). You could argue that you could save a few bucks by sticking with V.32 and going that bit slower, or relying on data comrpression (if you *know* that your transferred data will all be text), but it's best to know for sure that you won't kick yourself in the future for making the wrong decision. -- Geoffrey Welsh - Operator, Izot's Swamp BBS (FidoNet 1:221/171) root@zswamp.uucp or ..uunet!watmath!xenitec!zswamp!root 602-66 Mooregate Crescent, Kitchener, ON, N2M 5E6 Canada (519)741-9553 "He who claims to know everything can't possibly know much" -me