[ Home ] [ Table of Contents ] [ About Lee Goeller ] [ Search ]

The PBX Scene: July, 1986

Some Recent Events

GTE has been involved in various adventures during the last quarter. For one thing, GTE and Siemens started talking about a joint venture to hit the transmission and central office market, while GTE and Fujitsu have been working on a joint PBX marketing venture (GTE PBXs, not the Fujitsu Focus). On top of all this, GTE decided to stop making telephone sets, particularly those used on the Omni, and then relented on behalf of the FeatureCom VI. Whatever the big picture may be, the small pictures are sure moving fast.

In other mergers, Wang, on the heels of a story in Fortune (Feb 3) about its various difficulties, has bought the rest of InteCom. Whether buying further into a capital intensive business like PBX manufacture in an overloaded market is the way to fight a slump is anybody's guess, and InteCom's loss of a law suit to American Network, where the latter charged, among other things, fraud, and the jury agreed, cannot do much good. Perhaps Wang's marketing of their smaller Telenova PBX as part of a packaged data communication and office automation system will set things straight. But Wang will have to keep moving in the office automation market. After all, Sony has bought into the office humanation of CXC.

From a general business point of view, all the above attracted far less attention than Burroughs eating Sperry.

The Rolm Redwood

"It's a major product announcement," the young lady said over the phone. "I have never heard of a minor product announcement," I responded. She wouldn't say what IT was, but I found out in Atlanta, at ICA: Redwood, Rolm's new PBX. As Susan Kutner said proudly, "It can be either a key system or a PBX, but it's not a hybrid." The supposed advantage is that, as a key system, CO trunks cost much less in some areas than identical facilities used as PBX trunks.

The whole thing depends on what one means by Key, Hybrid and PBX; one must approach the subject with care. It should be noted that all "electronic key systems" and "hybrid systems" that I have ever studied are PBXs, period. Trunks from the CO terminate on trunk cards which reach lines to telephones via a switching matrix. The matrix is controlled by one or more computers, and the telephone sets, which terminate in line cards, can be plain sets or sets with multiple buttons on them. If it looks like a duck, quacks like a duck, and swims like a...

What seems to make the difference to the non-technical regulators is that a trunk can "appear" on a button on one or more sets, and the station user can reach that trunk by pushing the button. Lamping, similar to that found in 1A2, shows which trunks are busy, idle, ringing, on hold, etc. What eludes our regulatory friends is that pushing a button to select a trunk sends signaling information to the central processor which then makes the line-trunk connection. How is this different from dialing 9 or some other access code to the central processor which then makes the line-trunk connection? I marvel at the ability of political appointees, ably assisted by their staff lawyers and economists, to make such subtle distinctions.

The Redwood boasts of being "fully digital." The obvious question, unanswered, so far as I could tell, in Rolm's fat press kit, is "What KIND of digital?" Is it the traditional 12x12 (12000 samples per second coded into a 12 bit linear word) that Rolm has used in the CBX for years, or is it 8x8, the format of T-Carrier, ISDN, etc? Or is it delta mod like that found in IBM's SBS? I finally called Susan Kutner, and she admitted, wonder of wonders, that Rolm has finally caught up with the rest of the industry and is using 8x8. Wow! Can this possibly suggest something for the future in the CBX? Maybe CBX III? Let's hope.

After all, Redwood sets are Rolmphones connected to line cards by Rolmlink, just as in CBX II. They use standard T-Carrier 64 Kbps and, in the Redwood line cards, do not need to be mapped into the earlier 12x12. This has to be an advantage. But Rolmlink provides a 256 Kbps bandwidth in each direction (on a single pair). The intent is to use 64 for voice, 64 for data, and have two more channels available for signaling, sync, etc. But Redwood does not, as yet, handle data. Because a matrix path in Redwood only requires one third the bandwidth of a 12x12 path, the need to sub-multiplex, used in CBX II, would seem to be unnecessary; in Redwood, it should be possible to circuit-switch voice and data alike, using standard 64 Kbps matrix paths, since voice and data look alike coming from the set. With this NOT being done, one cannot help but wonder if there aren't enough matrix paths, or if some other factor blocks the data stream but not the voice stream. Stay tuned. This one ought to be most interesting.

On the way to the Fourth Generation

In the past, I have suggested that a fourth generation PBX would be one designed from the beginning to handle voice, data and image. Adding image, which at present takes at least an order of magnitude more bandwidth than voice or data, seemed to be a challenge that would require either fiber optics to each user terminal, or incredible sophistication in image processing to permit use of something as meager as a T-carrier 64 Kbps channel from switch to set. Either way, maybe ten years.

Well, the French are on their way. In Biarritz, the whole town is wired with fiber and a "Videophone" system is in operation. Not only can you call up your friends and see their lips move in real time, you can call a central library and watch the movie of your choice on your telephone. The videophone set, made by Thomson, goes for about $3000, while the fiber optics comes to about $7000 per home. Fortunately, French Government is picking up the tab. Three other cities, including Paris, are wired up and ready to go (Times, April 21, 1986, page D12). It would be interesting to know how all this investment is being justified.

If the French have become converts to Gebhardt's law (you can make anything run on pure money), good old Yankee ingenuity, Texas style, may offer the cost conscious a more reasonable alternative. A company called Image Data Corporation, of San Antonio, TX, is galloping across the country demonstrating their "Photophone." The Photophone is a slow-scan TV system that sends still pictures over ordinary dial-up telephone lines. Colorado Video, RCA and others have had similar systems for 15 years or so, but not like this. I thought I had seen everything until I saw the Photophone demo.

The Photophone uses a conventional TV camera to catch the image, but it digitizes and compresses the signal immediately. The compression algorithm is beyond belief; the compressed, digitized picture is packed into a series of packets complete with error correcting capability, and the packets can be zapped over a dial-up telephone line IN FIVE SECONDS! In addition to the compression algorithm, Image Data has built its own 9.6 Kbps modem to match the signal. If a TV picture (500x400 lines) can be sent in 5 seconds at 9.6, the picture must be about 50 kilobits. This is the equivalent 4 or 5 typed pages, double spaced.

Because of the coding and packetizing, the picture is unaffected by line loss, phase shift, moderate noise, etc. Send a picture to another Photophone, then return it, and you can't tell the returned picture from the original. And the pictures, although black and white, are GOOD. Put a book in front of the camera, catch a page with both text and photographs, and send it off. Both text and pictures come out like the original. With the short transmission time, a speaker can select a diagram and send it while he pauses for breath; you don't need a separate line for voice. Great way to do teleconferencing, lecturing to audiences at several locations, etc. A "composite video" output is available to drive external monitors in addition to the 9" tube built in, and a composite video input can be accepted from a computer, for instance, so that you can intersperse your pictures with spread sheets or whatever. A built-in floppy disk lets you capture pictures from either end, or view or send stored pix.

The Photophone costs about $8600 per unit, almost the same as the analog Colorado Video equipment. But it is much faster, and the picture quality seems to me to be superior.

The Photophone is designed for use in conference rooms, colleges, etc., where remote presentations of very high quality are required. But for residential use, where a fixed-focus "talking head" is desired at the far end, the "Luma," from Luma Telecom, a California company, may be quite adequate, particularly at a suggested price of $1500. I haven't seen this one, but it, too, is reported to have a 5 second transmission time, and a pretty good picture on a 3" screen.

The problem with all this is that the picture people are getting way ahead of the PBX people. They are handling their image information so effectively that they do not need a 64 Kbps channel or digital switching. It looks like I'll have to have a new definition of a fourth generation PBX.

What's in this Mailing

The new system this time is the Tadiran Coral, a voice-data digital PBX that covers the small end of the market. Updates include the Redcom and Solid State Systems's Jr. Exec; both companies have added proprietary phones, and both have brand new factories. They must be doing something right!

[ Top ] [ Next ] [ Table of Contents ]


Copyright 2005 Lee Goeller. All Rights Reserved.