Proposed Metrics for TGT and Call to Action Month Year doc.: IEEE 802.11-00/1202r0 October 2004 Proposed Metrics for TGT and Call to Action Date: Oct 14, 2004 Author: Charles R. Wright Azimuth Systems Acton, MA Ph: 978-268-9202 charles_wright@azimuthsystems.com C. Wright, Azimuth Systems C. Wright, Azimuth Systems
October 2004 Introduction We’ve talked in a lot of generalities about aspects of wireless performance that impact the user experience These can be translated into specific metrics With corresponding test environments, etc. We need to get down to the business of Defining the metric Documenting the measurement methodology This presentation tries to provide a condensed list of metrics C. Wright, Azimuth Systems
Presentations driving this one October 2004 Presentations driving this one 11-04/987r0, “A First Stab at 802.11 Metrics,” Mandeville 11-04/989r1, “Metrics for Characterizing BSS Transition Time Performance,” Wright, Polanec 11-04/1009r1, “Framework, Usages, Metrics Proposal for TGT,” Mehta, et al 11-04/1017, “Comments on Wireless Performance & Prediction Metrics,” Kobayashi, et al 11-04/1157r0, “A Metrics and Methodology Starting Point for TGT,” Wright 11-04/1156r1, “Bottom-Up Evaluation of 802.11 Performance Testing,” Foegelle C. Wright, Azimuth Systems
What metrics need to be defined? October 2004 What metrics need to be defined? Several categories Link layer Physical layer Antenna performance 802.11 link management related Whole device Two kinds of equipment Access points Client stations Environment types Define environments that are appropriate to the specific metric C. Wright, Azimuth Systems
October 2004 Link Layer Metrics Goal of these metrics is to measure device performance under optimum conditions “Line of sight” – no multipath Conducted tests to avoid interference problems Need to specify security, QoS features in use “Throughput” Max forwarding rate, forwarding rate at max offered load (FRMOL) TCP data throughput MSDU loss Loss after wireless link retries A function of offered load or only at maximums? Delay (latency) Definitely applicable to APs Client applicability? Measurability? Jitter Same comment as for delay “Rate versus Range” Throughput versus input signal level C. Wright, Azimuth Systems
Physical Layer Metrics October 2004 Physical Layer Metrics Receiver sensitivity at each PHY rate Definitely measured without multipath (“line-of-sight”) Probably measured with multipath, too Adjacent channel interference at each PHY rate Probably not measured with multipath Diversity performance Waiting for an expert to consider such a test Applicable to all 802.11 devices Questions Other PHY layer metrics? Which multipath models? Are these tests only performable using “radio control” software? C. Wright, Azimuth Systems
Antenna Performance Single antenna pattern October 2004 Antenna Performance Single antenna pattern One for each antenna? Applicable to all 802.11 devices Questions How do we deal with antenna loading caused by humans or other stuff? How do we characterize “smart” antennas? This slide has too much white space on it! C. Wright, Azimuth Systems
802.11 Link Management Related October 2004 802.11 Link Management Related AP Association Capacity, Association Rate How many clients can an AP support? How fast does the AP associate them? Variables: authentication method Transition time metric How fast does “roaming” algorithm work? Likely we should wait until TGr is more solid Is it a station, AP or system test? Other metrics related to TGk We need to investigate this – I have no idea what metrics might be needed, if any C. Wright, Azimuth Systems
October 2004 Whole Device Tests Don’t know if we actually want or need this, but it’s here so we don’t forget it Rate vs. Range test done here too? Anechoic Chamber 802.11 Device Bidirectional Multipath Simulator DUT Adjacent Channel Intereferer V.A. = RF signal path Traffic Generator & Analyzer Ethernet C. Wright, Azimuth Systems
Conclusions There is quite a list of metrics! October 2004 Conclusions There is quite a list of metrics! Let’s get people to take responsibility for some of them Want to see proposals or at least discussions of all these metrics during the 14 hours of meeting time in San Antonio C. Wright, Azimuth Systems