> -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of Mustafa Hasan > Sent: Tuesday, July 15, 2008 8:08 AM > To: Adams Jon; [EMAIL PROTECTED] > Cc: [email protected] > Subject: Re: [6lowpan] Regarding coexistence with 802.11n > > Dear Jon & Geoff, > > Thank you very much for such a nice and detail mail.... It helped me a > lot. I want to add few comments regarding this issues, after my little > research in this field. > > 1. Even thr are some free channels available for different region, but > still creating problem for IEEE802.15.4 traffic. Actually, high power > of wifi is killing ieee802.15.4 traffic (testbed results). > > 2. The current research on this issue is interestingly biased, some > are saying its not a big problem (e.g ZigBee Alliance) and others > (e.g. Z-Wave Alliance) are saying its a big issue.
The 802 family of standards has some modicum of coexistence "built-in" as Jon suggested - our experience is that a serious interference source are non-802 devices including various FHSS devices that pay little to no heed to whatever else is in the band. The often have no coexistence considerations and blatantly use the band as if they were the only device in town. Some of these devices are getting smarter though (but for their own selfish reasons). Various first generation 2.4GHz cordless phones used to horrible spectrum users by using the entire 2.4GHz band with little regard for anything else. They seem to be taking some steps now to segment the band and avoid Wi-Fi sized areas now; probably not to help Wi-Fi but to make themselves work better by avoiding sections of the band that don't work well for *them*. > 3. Distance between ieee802.15.4 and ieee802.11x is a big problem in > this case. If I summarize all the research efforts (ieee802.19 > coexistence assurance documents, white paper(ZigBee, Z-Wave) and many > research papers), its not possible that 15.4 & 11x stay together in a > distance of <2 meters (even tried with offset 20MHz). Our practical experience is 2 meters is plenty in most situations, even 1 meter is enough but getting riskier. A lot of the vendor/organization white papers seem incredibly biased to me and upon reading them, I basically get the idea that the network running right on my desk couldn't possibly work (and might even kill me ;) when clearly it is working. > 4. thr are many potential application require this distance to be in > few centimeters. May be some technique in upper layers can give little > bit reliability... but is it enough??? > > I am looking for some help, comments and suggestion, how coexistence > problem can be solved for very short distance scenario... For a few centimeters, this problem may very well be unsolvable (at reasonable cost/effort). The only way the upper layers can accommodate something as brutal as this is through coordination and knowing that the radio will effectively be "blinded" at specific times. For uncoordinated networks, this isn't really practical. At centimeter distances, *everything* can act like an antenna and it is very easy to saturate a radio receiver which effectively blinds it from receiving anything - as you pointed out, even significant frequency separation isn't enough when things are closely located; there is just so far to run in a relatively small band. Even closely located *receivers* can interfere with each other even if they never transmit; we have had to use shielding between closely colocated receivers in some multi-receiver designs to prevent them from interfering with each other. In simple terms, if it oscillates, it radiates and something either sensitive enough or close enough will get effected by such radiation possibly to the point of incorrect operation. Few centimeter separation of transmitters is a difficult problem to solve and would likely be out of scope of this effort. -- Christian Herzog Software Technologies Group, Inc. mailto:[EMAIL PROTECTED] http://www.stg.com +1 (708) 547-0110 x225 FAX +1 (708) 547-0783 _______________________________________________ 6lowpan mailing list [email protected] https://www.ietf.org/mailman/listinfo/6lowpan
