Re: [802.19] Coexistence Metrics

Thread Links	Date Links
Thread Prev	Thread Next	Thread Index	Date Prev	Date Next	Date Index

Mariana,

Thanks for your comments. Sorry I did not get to them yesterday but it was a busy day. I see that Paul had some questions and I guess I have some too.

See comments/questions below.

Steve

From: Mariana Goldhamer [mailto:marianna.goldhammer@alvarion.com]
Sent: Thursday, September 06, 2007 7:16 AM
To: Shellhammer, Steve; stds-802-19@ieee.org; Adrian.P.Stephens@intel.com; aghasemi@crc.ca; bji@sta.samsung.com; bkraemer@ieee.org; dlubar@ieee.org; david.grandblaise@motorola.com; dickroy@alum.mit.edu; dougchan@cisco.com; eldad.perahia@intel.com; fm@octoscope.com; john.sydor@crc.ca; ksohrabi@nextwave.com; KStanwood@cygnuscom.com; Naftali Chayat; Nat.Natarajan@motorola.com; ppiggin@nextwave.com; Shlomo Malka; Shyamal.Ramachandran@motorola.com; wuxuyong@huawei.com; Ziv Nuss
Subject: RE: Coexistence Metrics

Hi Steve,

I've spoke with Alvarion's people doing deployment simulations and we agreed that the following should be added:

The outage probability for the elements of the 802.16 network, for the case that 802.16 network has uniformly spread element positions and the 802.11 network

has variable element positions; the reason for fixing the 802.16 network elements is due to the higher "foot print" resulting from OFDMA sensitivities

SJS> When you say “network elements” are you referring to the base station or the client? I can understand having a planned (uniformly spread) deployment of base stations but I cannot understand such a deployment of clients. The users get to use the clients where ever they want.

SJS> I am fine with using outage probability as a coexistence metric. To do that we need to define what we mean by an outage. How would you like to define an “outage.” One idea is to specify a maximum allowed packet error rate on the uplink or downlink to that client and if that is exceeded the link is in outage.

The percentage of position instances for which the outage of each element is less than 2%

SJS> If we use fixed positions this is straightforward. If we place the stations randomly then we need to decide how close two placements need to be to count them as the same position.

The statistics of the distance between any of the elements of 802.16 network and any of the elements of the 802.11 for those instances enjoying less than 2% outage

SJS> This is an interesting idea. So we identify when we have a 2% outage (actually I think I would include the 2% in the definition of an outage) and look at the statistics of that node and probably the closes node from the other network.

If required, we can change to have 802.11 as the fixed network.

SJS> My preference is for a random placement of stations.

The "outage" may be evaluated using C/(N+I) criteria for every element.

SJS> I think that SINR can be used to determine if there is a packet error. My personal preference is related to Packet Error Rate since to me that is what matters, did the packet get through.

Mariana

From: Shellhammer, Steve [mailto:sshellha@qualcomm.com]
Sent: Thursday, September 06, 2007 8:48 AM
To: stds-802-19@ieee.org; Adrian.P.Stephens@intel.com; aghasemi@crc.ca; bji@sta.samsung.com; bkraemer@ieee.org; dlubar@ieee.org; david.grandblaise@motorola.com; dickroy@alum.mit.edu; dougchan@cisco.com; eldad.perahia@intel.com; fm@octoscope.com; john.sydor@crc.ca; ksohrabi@nextwave.com; KStanwood@cygnuscom.com; Mariana Goldhamer; Naftali Chayat; Nat.Natarajan@motorola.com; ppiggin@nextwave.com; Shellhammer, Steve; Shlomo Malka; Shyamal.Ramachandran@motorola.com; wuxuyong@huawei.com; Ziv Nuss
Subject: Coexistence Metrics

All,

Here is a first cut at a document on coexistence metrics for the 3650 MHz band. We can discuss it on the call and I will take suggestions on any additional coexistence metrics.

Ian, please post on the web.

Steve