TELECOM Digest Mon, 10 Jan 94 01:43:00 CST Volume 14 : Issue 23 Inside This Issue: Editor: Patrick A. Townson Re: Multi-line BBS's (Lars Poulsen) Re: Multi-line BBS's (Fred R. Goldstein) Re: Correction: Re: Help Needed With V.42bis (ssatchell@bix.com) Re: Merlin Question (Christopher Zguris) Re: US Digital Cellular Standard (Donald J. Miller) Re: 500 Channel Cable Television (David M. Berman) Re: 500 Channel Cable Television (sandyron@delphi.com) Re: Notice to AT&T Long Distance Customers (Laurence Chiu) Re: User Interface From Hell (Gary Bridgewater) Re: "Dynamic" SLIP (John Kennedy) TELECOM Digest is an electronic journal devoted mostly but not exclusively to telecommunications topics. It is circulated anywhere there is email, in addition to various telecom forums on a variety of public service systems and networks including Compuserve and GEnie. Subscriptions are available at no charge to qualified organizations and individual readers. Write and tell us how you qualify: * telecom-request@eecs.nwu.edu * The Digest is compilation-copyrighted by Patrick Townson Associates of Skokie, Illinois USA. We provide telecom consultation services and long distance resale services including calling cards and 800 numbers. To reach us: Post Office Box 1570, Chicago, IL 60690 or by phone at 708-329-0571 and fax at 708-329-0572. Email: ptownson@townson.com. ** Article submission address only: telecom@eecs.nwu.edu ** Our archives are located at lcs.mit.edu and are available by using anonymous ftp. The archives can also be accessed using our email information service. For a copy of a helpful file explaining how to use the information service, just ask. TELECOM Digest is gatewayed to Usenet where it appears as the moderated newsgroup comp.dcom.telecom. It has no connection with the unmoderated Usenet newsgroup comp.dcom.telecom.tech whose mailing list "Telecom-Tech Digest" shares archives resources at lcs.mit.edu for the convenience of users. Please *DO NOT* cross post articles between the groups. All opinions expressed herein are deemed to be those of the author. Any organizations listed are for identification purposes only and messages should not be considered any official expression by the organization. ---------------------------------------------------------------------- From: lars@Eskimo.CPH.CMC.COM (Lars Poulsen) Subject: Re: Multi-line BBS's Organization: CMC Network Products, Copenhagen DENMARK Date: Sun, 9 Jan 94 23:30:05 GMT In article dannie@coplex.coplex.com (Dannie Gregoire) writes: > I would like to know how some of these bulletin boards have 60-100 > lines running into them (eg EXEC-PC). Do they simply have that many > individual lines run or is there a nifty service that the TELCO offers > through a PBX? A BBS is just a special case of the general class of "multi-user computer systems". Access for remote terminal users of such a system can be provisioned in several ways: 1) If the users are on-site, the system may be constructed with many serial ports, and you may wire each terminal directly to a port dedicated to that terminal. The drawback of this is, that for large systems, serial terminal ports may actually be very expensive (as expensive as high-speed ports, and probably more expensive than the terminals.) 2) Often the terminals are wired to "cluster controllers" that aggregate traffic from a number of terminals (16 is a common number) and feed the aggregated traffic into a high-speed port. The operating system on the server machine must then contain code to unravel the multiplexed data before feeding it into the software driver for the logical ports. 3) If not all of the users are active at the same time, it has been common to attach the terminal to a modem, and attach the modem to a port on the building's PBX, which then routes the calls to a bank of modems which may be attached to ports on the server (as in 1) or on a cluster controller (as in 2). This allow you to get by with a smaller number of ports, and also allows all wiring to be installed by telephone technicians, which traditionally have been easier to find and manage than computer technicians. Also, the PBX already has mechanisms to deal with contention. 4) This technique extends in a simple manner to off-site users: Just attach a few modems to outside lines instead of PBX station ports. You need one outside line with a modem per port, of course. 5) If the callers are far away, you may be able to save by locating the cluster controllers near the users, far from the server system, and attach them to the server by: - leased data lines or - automatic dialers that bring the line up when someone needs service. Generally, if you have a remote building with a cluster, it will be more economical to lease the line. The crossover happens at 2-4 hours per day. If you have 4 to 8 terminals, then it is likely that at any given time at least one will be in use. 6) Some companies have specialized in setting up such remote cluster controllers (called PAD: Packet Assembler/Disassembler) provide them with local dial-in lines with modems, and arranging for shared carriage from these access points to servers in several/many cities over lines leased from the phone companies. These companies (legally called "Value Added Networks") generally use a protocol called X.25 between the cluster controllers and the server hosts. They also provide for connections directly between the servers using the X.25 protocol. The VAN carriers include CompuServe, SprintNet (formerly TELENET), TYMNET, AT&T Accunet, INFONET and many more. With such a hookup, you may be able to use a single high-speed access line to carry traffic originating at hundreds of different remote points. 7) Finally, a whole other group of carriers have sprung up to do the same using a more modern set of protocols called TCP/IP or Internet Protocol. Generally, this is much less expensive than X.25 service, and provides many more features. Principal commercial Internet service providers who offer service throughout the USA include: ANS (Advanced Networks and Services - a partnership including IBM, MCI and the State of Michigan), AlterNET (a service mark of UUNET Technologies), PSI (Performance Systems International of Troy, NY), AT&T, INFONET and SprintNet. (At least the last 3 also offer one-stop international connections.) The special magic of this technology is that the networks are all interconnected, so that your customers can reach you even if they have service from different providers. The company that I work for, is on of many that sells equipment for Internet service. We expect (between all of us) to hook up every school and every public library the US and at least Western Europe in my lifetime. We have a long way to go. To date we have only connected about two million computers, serving about ten million people worldwide. Lars Poulsen Internet E-mail: lars@CMC.COM CMC Network Products Phone: (011-) +45-31 49 81 08 Hvidovre Strandvej 72 B Telefax: +45-31 49 83 08 DK-2650 Hvidovre, DENMARK Internets: designed and built while you wait ------------------------------ From: goldstein@carafe.tay2.dec.com (Fred R. Goldstein) Subject: Re: Multi-line BBS's Date: 9 Jan 1994 03:18:18 GMT Organization: Digital Equipment Corp., Littleton MA USA In article dannie@coplex.coplex.com (Dannie Gregoire) writes: > Hi Pat: > I'll direct this question to you if possible, as you are the true > phone system guru. I asked it in the newsgroup a couple of months > back with no useful response. I would like to know how some of these > bulletin boards have 60-100 lines running into them (eg EXEC-PC). Do > they simply have that many individual lines run or is there a nifty > service that the TELCO offers through a PBX? I apologize if this is > a stupid question, but it is one that has baffled me, and I gotta know > the answer. Thanks for any help ... > [TELECOM Digest Editor's Note: Thanks for the compliment, but you > overestimate my skills a little. Depending on your application or > needs, you can have as many actual lines run as desired. I suspect > most very large systems these days however use what is called T-1 > or similar, where a large number of circuits are multiplexed or > handled over just a few actual pairs of wires. In addition to T-1, > there are similar methods for bringing in a large number of circuits > on only a few wires. In my own personal applications in the past, > I always just had the physical wires, but that was several years > ago before the present technology became available. Perhaps Fred > Goldstein or one of the *real* tech people here will reply. PAT] I have to admit that I don't know much about how real BBSs operate, just how they could be operated. But thanks for the compliment, Pat. BBSs are funny things. I read an article about Channel 1, a huge operation in Cambridge, which services its dozens of incoming lines by a network of dozens of little PCs. They were lined up all over a tiny house. If it were up to me, I'd have a big machine or two ... It's often easier and cheaper to use T1 circuits rather than the 23-24 (30 in Europe) separate analog circuits that it replaces. If your central office is digital (eg., 5ESS or DMS), then a T1 is much easier for the phone company to provide -- it plugs right in to a trunk port. Most large PBXs probably hook up this way now. For modems, though, this isn't always cheaper. You need either a T1 modem (big bucks for a BBS operator; figure $1k/line, from Primary Access or US Robotics) or a channel bank (maybe $5k, but available used) to demulitplex it into analog lines. Of course you could use a PBX instead but a plain old channel bank is lots cheaper since it leaves out the switching function! So I suspect the BBS crowd tends towards big blocks of analog lines. Advantage of moving towards T1: If you're a plain old "kiddiecomms" BBS, then maybe this won't be necessary. But if you use T1, you can not only talk to modems, but talk to ISDN and Switched 56 at a 56 kbps speed. Channel banks, PBXs and T1-modems all support this. Fred R. Goldstein k1io goldstein@carafe.tay2.dec.com Opinions are mine alone; sharing requires permission ------------------------------ From: ssatchell@BIX.com (ssatchell on BIX) Subject: Re: Correction: Re: Help Needed With V.42bis Date: 8 Jan 94 18:57:34 GMT Organization: Delphi Internet Services Corporation Let me add my pair-o'-pennies to the V.42 bis discussion. Much of what I'm about to say is derived by my research into the question for inclusion into TIA PN2826, what will eventually become TSB-38, a modem testing methodoly description for modems: BTLZ is not a true LZW derivitive, as jim claims. (However, the difference is so minimial that only the truly nitpicking will care.) It is Ziv-Lempel with several additions to permit adaption and to prevent expansion of the data, particularly pre-compressed data. It also has one of the niftiest escape systems I've ever seen. You start with a dictionary pre-filled with the alphabet, and have nine-bit output codes. As data transfer progresses, and the dictionary fills, you expand the size of the output code until you reach the dictionary limit negotiated by both ends. At that point, you start "burning" leaf nodes so that as the characteristics of the data change the LZ tree will change shape to adapt. If a block of compressed data is significantly larger than its uncompressed counterpart, then you enter "transparency mode" and send the uncompressed data. If a subsequent block of compressed data is significantly smaller than its uncompressed counterpart, you then switch to "compression mode". Initally the system is in transparent mode. While in transparent mode, if an escape character is seen, the system sends an escape-in-data sequence like most codes which use a reserved flag character. Unlike the others, though, the escape character definition is CHANGED so that repeated runs of the (original) escape character don't cause a two-fold explosion in the size of the resulting data stream, and a less-than-optimal change to compression mode for that string. For PN2826, we wanted to develop test files which would virtually guarantee one million data bits sent through the signal converter (before HDLC zero- stuffing). So we examines a number of files to determine the compression properties using dictionary sizes common to modems. What I found is that the vast majority of uncompressed files experience a compression ratio (input/output) ranging from 2.5:1 to 3.1:1; the only time I got anywhere near the four-to-one everyone attributes to V.42 bis is when I was transferring B&W line-art in uncompressed TIFF form. (Indeed, with one picture I achieved a 7:1 compression ratio.) Compressed files had ratios which depended greatly on the scheme used to com- press them. LZW-based compressed files showed anywhere from 1.1:1 to 1.4:1 with V.42 bis; a purely random file (generated using the DES algorithm) has a compression ratio of 0.998:1 -- the expansion attributed to the 57 escape-in- data sequences in this 132-kilobyte file. Encrypted compressed files tended to be at exactly 1:1. These figures were developed by running files through a desktop implementation of V.42 bis -- my thanks to Dr. Coleman of Georgia University for the code. This is a straightforward implementation of V.42 bis with a rather wasteful compression prediction algorithm (fill two buckets, throw out the one that's more full) but is simple and surprisingly good at minimizing local expansion. ------------------------------ Date: Sun, 9 Jan 94 15:02 EST From: Christopher Zguris <0004854540@mcimail.com> Subject: Re: Merlin Question In a recent TELECOM Digest post Vince Dugar (vdugar@stortek.stortek. com) wrote: > Why does the Merlin system charge users so much (I forget now, but > it's a lot) to buy a special modem adapter? Is there a cheaper > solution? What about using an acoustic modem? (only want it for > CompuServe mail handling, so low baud would be OK) The simplest way I know of - that I have used successfully - is to tap the voice pair on the merlin. As I remember, the voice pair on the merlin are the two middle wires on the 8 pin jack. If you're electronically inclined (or know someone who is) the simple schematic below should do the trick. To build it, you'll need two identical RJ-45 8-position 8-conductor phone jacks, and another merlin cable. To build it, wire pins 1-3 and 6-8 straight through from the "from Wall" jack to the "to phone". Tap pins 4 & 5 with a cable to plug into the modem (or run them to another jack and plug the modem in). The SPST switch is optional, but will come in handy, read on. To use the thing, plug the cord coming from the wall in the "from wall" jack and plug your new cord into the "to phone" jack, then plug that cord into your phone. You're phone should work fine, if it doesn't, disconnect the gizmo and check you're wiring. To actually use the thing, fire up your modem, dial out, and QUICKLY pick up on a line USING the merlin phone. When the modem dials it will dial out through the merlin on whatever line your merlin has selected! The disadvantage to this gizmo over the AT&T rip-off is that you must keep the merlin off-hook while your modem is in use and manually hang up when your modem is done (redialling is impossible- if you want to do that manually dial and redial using the merlin). You will want to set the SPST switch to open so the voice pair disconnected from the merlin, otherwise the modem noise will come through the merlin speaker or handset, and if the handset it used voices and noises will be picked up and sent and screw up your modem communications. I've used a similar setup at 2400 baud, anything above that I can't vouch for. When you're done, flip the SPST switch. If you have any questions let me know, I'll try to help out- I love solving problems! {From Wall} 1 2 3 4 5 6 7 8 [;][;][;][;][;][;][;][;] 8 Position-8 Cond. RJ-45 Jack I I I I I I I I I I I I *---------------------- \ I I TAP-wire to modem (red & Green) I *------------------- / I I I I I I-------------O / I / SPST Switch I I-------------O I I I I I I I I I I I I [;][;][;][;][;][;][;][;] 8 Position-8 Cond. RJ-45 Jack {To Phone} (First attempt at ASCII art, what do you think?) Chris ------------------------------ From: dmiller@crl.com (Donald J. Miller) Subject: Re: US Digital Cellular Standard Date: 9 Jan 1994 13:59:23 -0800 Organization: CRL Dialup Internet Access (415) 705-6060 [login: guest] Weiyun Yu (weiyun@extro.ucc.su.OZ.AU) wrote: > It has come to my attention that the digital cellular standards > adopted by US carriers are not going to be compatible with what we > have adopted in Australia, GSM. I am interested in finding out a bit > more about the US systems but cant find any FAQ on the subject. The US Digital Cellular scheme (TDMA) was originally conceived to ease the bandwidth requirement for a phone conversation by at least 3 to 1. A 6 to 1 capacity advantage over the regular AMPS service would be achieved if and when acceptable half-rate voice codecs were available. Currently, each caall uses two of the six TDMA time slots. The aim was/is to eventually use only one. The first TDMA phones were to be dual-mode: that is to say that they would function as regular AMPS phones as well. The hope WAS that after 10 to 15 years, the AMPS functionality could be dropped, resulting in more cost effective phones. A monkey wrench has been tossed in the works, however. AFTER the cellular industry chose TDMA as the standard, Qualcomm proposed the use of a different CDMA technology with promises of even greater capacity. Some, but not all carriers joined the Qualcomm camp. The net result is that we now have two digital phone standards. Motorola proposed a new analog system (NAMPS) with a capacity advantage of 3 to 1 over AMPS that many hail as a good intermediate step before full digital cellular implementation. So, now we have FOUR phone "standards". What about ROAMING? Well, it looks like the more expensive DUAL-MODE phones are here to stay. Either of the two digital systems, TDMA or CDMA could have been cost effective with time and further work on the ICs inside the phones. The power-wasting RF duplexers required in AMPS phones for full-duplex operation would not have been needed -- money would be saved and talk time increased. Six times the existing bandwidth was not enough. We got greedy. Don Miller dmiller@crl.com ------------------------------ From: images@netcom.com (David M. Berman) Subject: Re: 500 Channel Cable Television Organization: NETCOM On-line Communication Services (408 241-9760 guest) Date: Sun, 9 Jan 1994 18:46:22 GMT I think most of you are limiting yourselves when you imagine the offerings that could be available with video dial-tone or 5000 channels of audio/video/data. Imagine not only every piece of video, film, or music ever published, but the new publishing opportunities that will spring from recent advances in home/cheap video and audio production. You might see excellent products aimed at smaller and smaller audiences (college lectures, poems, paintings, dance, how to fix your 1983 Toyota, etc.). To me, the most exciting thing about the possibility of all this new bandwith is the thought that we could escape the tyranny of the majority and their pedestrian tastes. David M. Berman images@netcom.com ------------------------------ From: SANDYRON@delphi.com Subject: Re: 500 Channel Cable Television Date: Sun, 9 Jan 94 22:51:21 EST Organization: Delphi Internet I somewhat agree. Who has the most bandwidth in the home? Anyone of the 500 channels could be an on-demand channel. By the way, without a broadband entry in the house how can all of the consumers requests be met like pix in pix. ------------------------------ From: lchiu@crl.com (Laurence Chiu) Subject: Re: Notice to AT&T Long Distance Customers Date: 9 Jan 1994 20:24:23 -0800 Organization: CRL Dialup Internet Access, California Reply-To: lchiu@crl.com In article , William M. Eldridge wrote: >> According to a wire service account in the {Boston Globe}, AT&T is >> changing their rates to be more like MCI and Sprint. The list price > As somebody who just switched from AT&T to MCI, I have a few > qualifications for this. > On international calls, MCI has all weekend rates, while AT&T leaves > its three Day-Evening-Night slots the same, seven days a week. MCI > has better hours during the week. AT&T had worse setup (first minute) > charges. If you join Reach Out World from AT&T for $3 a month, then it seems that there are only two time slots and there is no higher first minute charge. It's quite competitive with MCI now, Friends and Family even taken into account. For example if I specifiy New Zealand as my Selected Country then during off peak times and weekends I get to call *any* number in NZ for 0.68/minute while with MCI and their Friends around the World plan and Friends and Family, I get to call only a maximum of two numbers for $0.66/minute. I find AT&T's plan more flexible here. Laurence Chiu | Walnut Creek, California Tel: 510-215-3730(wk) | Internet: lchiu@crl.com ------------------------------ From: gjb@lsil.com (Gary Bridgewater) Subject: Re: User Interface From Hell Date: 9 Jan 1994 08:41:39 GMT Organization: LSI Logic In article johnl@medusa.gsfc.nasa.gov (John Limpert) writes: > How can telephones be made easier to use? The local phone companies > are going to have a hard time selling new features to their customers > if they expect them to press "*-*-FLASH-4-2-#-6-6-6" every time they > use them. In another life, at a super-mini company, we had a subsidiary PBX design company that was designing a terminal/phone combo but the whole thing evaporated during a "rightsizing". I saw one working - real, not a prototype. There was a button on the terminal keypad that would bring up a phone menu and you could click on the option you wanted. Software on the computer could interact with the phone so you could have a Rolodex on-line and touch a number to dial it, record and playback digital messages and record and playback on-line (with a beep). You could then file, mail, etc. the messages. Took lots (and lots) of disk space but we sold disks so what the heck. Used realy cheap off-shore digital handsets - bad but liveable. Probably something like the Newton will solve this. Also, the VCR programming in another thread. Gary Bridgewater (gjb@lsil.com) LSI Logic, Milpitas, CA ------------------------------ From: warlock@csuchico.edu (John Kennedy) Subject: Re: "Dynamic" SLIP Date: 9 Jan 1994 18:52:35 GMT Organization: California State University, Chico In article , wrote: > My understanding of SLIP is that it is a point-to-point dedicated > configuration, requiring a modem on the receiving end to be dedicated > to a specific user (due to IP I think). If you get the right hardware, that isn't necessarily true. You could dedicate one line/SLIP connection if you wanted to, but many people don't. Most people dedicate one IP _addresses_ to a SLIP user. That mostly depends on how much you're willing to trust any IP address (for security purposes or otherwise ... like access to a USENET server). > I've heard some talk about so-called 'dynamic' SLIP -- where the SLIP > connection is made, but through a mux or terminal server, allowing the > provider to serve multiple dial-up customers instead of a 1-1 ratio. ... Yes, true. I use Cisco communication servers to do this, but there are many more vendors that provide the same features. Depending on your computer's OS, you might be able to connect a modem into it and server multiple SLIP users. The one thing I'd worry about is that most SLIP servers I've seen can only provide access to the subnet they are in ... so you can only server ~250 (fixed, single user/IP addr) users with one rotary. John Kennedy ; Communications Services; USENET admin ------------------------------ End of TELECOM Digest V14 #23 *****************************