Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!sdd.hp.com!decwrl!sgi!rpw3@rigden.wpd.sgi.com From: rpw3@rigden.wpd.sgi.com (Rob Warnock) Newsgroups: comp.protocols.tcp-ip Subject: Re: Mobile TCP/IP (was Re: Can subnets be separated by another net?) Message-ID: <64889@sgi.sgi.com> Date: 24 Jul 90 08:57:55 GMT References: <9007171938.aa18804@huey.udel.edu> Sender: news@sgi.sgi.com Reply-To: rpw3@sgi.com (Rob Warnock) Organization: Silicon Graphics, Inc., Mountain View, CA Lines: 43 In article <9007171938.aa18804@huey.udel.edu> Mills@udel writes: +--------------- | You may remember the ill-fated XTEN network that could be described as | an Ethernet radio operating at 10 GHz for metro-area coverage. Are we | now seeing XTENs in the sky? +--------------- Not exactly. The 10 GHz local radio was not so much an Ethernet as a whole bunch of Localtalks... ;-} ;-} XTEN was [supposed to be] closer to a "non-mobile digital cellular radio". While the carrier frequency used for local distribution was indeed 10 GHz, the data rate was a paltry 256 Kbits/sec, shared among all stations located in a "cell", a 6-mile (max.) radius quarter-circle from a "local node" (that is, an area of about 28 sq. mi.). [Cells could, of course, be smaller than that, with careful power budgeting and re-use planning, just like today's mobile cellular systems. But there were only four frequencies available for re-use planning.] The traffic from the four cells of each local node was to be gathered and sent to a "city node" (central office), via point-to-point links not part of the allocation at 10 GHz. City traffic was then to be TDMA'd via one or more satellite transponders at geo-sync. Thus, nothing moved around. The routers would have been virtually identical to today's IP routers, except maybe simpler since the physical plant was so much more hierarchically laid out -- everything not local to a city went up to a satellite and back. The total "peak-busy-hour" traffic for New York City in 1992 was estimated [in 1979] to be less than one FDDI's worth. And today a lot of people think FDDI is slow. (SONET, here we come! ;-} ) The only historically safe guess is that we'll continue to underestimate our networking bandwidth appetites... -Rob (p.s. Fall of 1979, my job at Xerox/XTEN was designing/evaluating multi-access protocols for the 10 GHz local radio...) ----- Rob Warnock, MS-9U/510 rpw3@sgi.com rpw3@pei.com Silicon Graphics, Inc. (415)335-1673 Protocol Engines, Inc. 2011 N. Shoreline Blvd. Mountain View, CA 94039-7311