1. QoS aware Load Balancing
Month Year
doc.: IEEE yy/xxxxr0
Nov. 2006
QoS aware Load Balancing
Date:
Authors:
Notice: This document has been prepared to assist IEEE It is offered as a basis for discussion and is not binding on the contributing individual(s) or organization(s). The material in this document is subject to change in form and content after further study. The contributor(s) reserve(s) the right to add, amend or withdraw material contained herein.
Release: The contributor grants a free, irrevocable license to the IEEE to incorporate material contained in this contribution, and any modifications thereof, in the creation of an IEEE Standards publication; to copyright in the IEEE’s name any IEEE Standards publication even though it may include portions of this contribution; and at the IEEE’s sole discretion to permit others to reproduce in whole or in part the resulting IEEE Standards publication. The contributor also acknowledges and accepts that this contribution may be made public by IEEE
Patent Policy and Procedures: The contributor is familiar with the IEEE 802 Patent Policy and Procedures < ieee802.org/guides/bylaws/sb-bylaws.pdf>, including the statement "IEEE standards may include the known use of patent(s), including patent applications, provided the IEEE receives assurance from the patent holder or applicant with respect to patents essential for compliance with both mandatory and optional portions of the standard." Early disclosure to the Working Group of patent information that might be relevant to the standard is essential to reduce the possibility for delays in the development process and increase the likelihood that the draft publication will be approved for publication. Please notify the Chair as early as possible, in written or electronic form, if patented technology (or technology under patent application) might be incorporated into a draft standard being developed within the IEEE Working Group. If you have questions, contact the IEEE Patent Committee Administrator at
Lei Du, DoCoMo Beijing Labs
John Doe, Some Company

2. Month Year
doc.: IEEE yy/xxxxr0
Nov. 2006
Abstract
In this presentation we proposed a QoS-aware load balancing strategy to improve wireless network throughput and QoS effectiveness in accordance with REQ2030 in TGv.
Lei Du, DoCoMo Beijing Labs
John Doe, Some Company

3. Motivation & Solutions
Month Year
doc.: IEEE yy/xxxxr0
Nov. 2006
Motivation & Solutions
Motivation
In addition to the existing element “BSS Load” including STA Station Count and Channel Utilization, we like to have more precise actual load status in current QBSS/BSS.
STAs corresponding to different Access Categories may request specific load information for some ACs for QoS awareness.
Solutions
Define “AC” Station Count to describe the detail AC constitution of STAs currently associated with the QBSS/BSS.
Define “AC” Channel Utilization to describe the frequency of channel occupancy arising from each AC traffic.
These IEs shall be helpful for roaming STAs to select an AP for better QoS effectiveness.
Lei Du, DoCoMo Beijing Labs
John Doe, Some Company

4. Beacon Frame Nov. 2006 Table 8—Beacon frame body Order Information
Month Year
doc.: IEEE yy/xxxxr0
Nov. 2006
Beacon Frame
Add two new elements in table 8:
Table 8—Beacon frame body
Order
Information
Notes 30
AC Station Count
The AC Station Count element shall be present in a QoS AP where dot11RadioMeasurementEnabled is true. 31
AC Channel Utilization
The AC Channel Utilization element shall be present in a QoS AP where dot11RadioMeasurementEnabled is true.
Lei Du, DoCoMo Beijing Labs
John Doe, Some Company

5. Probe Response Frame Nov. 2006 Table 15—Probe Response frame body
Month Year
doc.: IEEE yy/xxxxr0
Nov. 2006
Probe Response Frame
Add two new elements in table 15:
Table 15—Probe Response frame body
Order
Information
Notes 29
AC Station Count
The AC Station Count element shall be present in a QoS AP where dot11RadioMeasurementEnabled is true. 30
AC Channel Utilization
The AC Channel Utilization element shall be present in a QoS AP where dot11RadioMeasurementEnabled is true.
Lei Du, DoCoMo Beijing Labs
John Doe, Some Company

6. STA Statistics Report IE (1)
Month Year
doc.: IEEE yy/xxxxr0
Nov. 2006
STA Statistics Report IE (1)
Change the order 10 in table 31b as follows:
Table 31b－ Group Identity for a STA Statistics Report (continued)
Group Identity Requested
Statistics Returned 10
dot11BSSAverageAccessDelay Group (only available at an AP):
dot11STAStatisticsAPAverageAccessDelay (INTEGER),
dot11STAStatisticsAverageAccessDelayBestEffort (INTEGER),
dot11STAStatisticsAverageAccessDelayBackGround (INTEGER),
dot11STAStatisticsAverageAccessDelayVideo (INTEGER),
dot11STAStatisticsAverageAccessDelayVoice (INTEGER),
dot11STAStatisticsACStationCount (INTEGER),
dot11STAStatisticsACChannelUtilization (INTEGER)
11－255
None
Lei Du, DoCoMo Beijing Labs
John Doe, Some Company

7. STA Statistics Report IE (2)
Change Figure 85k as follows:
Month Year
doc.: IEEE yy/xxxxr0
Nov. 2006
STA Statistics Report IE (2)
Change the order 10 in table 31b as follows:
dot11STA StatisticsAP AverageAccess Delay
dot11STA Statistics AverageAccess DelayBestEffort
dot11STA Statistics AverageAccess DelayBackGround
dot11STA Statistics AverageAccess DelayVideo
dot11STA Statistics AverageAccess DelayVoice
dot11STA Statistics ACStation Count
dot11STA Statistics ACChannel Utilization 1
3-11
13-7
Octets:
Figure 85k---Measurement Report field format for dot11BSSAverageAccessDelay Group
Lei Du, DoCoMo Beijing Labs
John Doe, Some Company

8. Neighbor Report IE Nov. 2006 Table 43B—Optional Extensions
Change Figure 85k as follows:
Month Year
doc.: IEEE yy/xxxxr0
Nov. 2006
Neighbor Report IE
Change the table 43B in v Draft 0.05 as follows:
Table 43B—Optional Extensions
Sub-Element ID
Name
Reserved 1
TSF Information 2
Roaming Candidate Preference 3
QBSS Available Admission Capacity 4
AC Station Count 5
AC Channel Utilization
4 6-10 11
QBSS Load
12-255
Lei Du, DoCoMo Beijing Labs
John Doe, Some Company

9. Access Category Station Count IE
Month Year
doc.: IEEE yy/xxxxr0
Nov. 2006
Access Category Station Count IE
Element ID
Length
Station Count Bitmask
Station Count List
Octets: 1
2 x (total number of non-zero bits in Station Count Bitmask)
Figure X1--- AC Station Count element format
Definition:
- The Station Count Bitmask field indicates the AC values that have Station Count specified in the following Station Count List.
- The Station Count List comprises a sequence of Station Count fields corresponding respectively to the non-zero bits in the Station Count Bitmask field.
- The Station Count field specifies the number of STAs that are currently associated with this QBSS for the corresponding AC traffic.
Table 44a－ Station Count Bitmask Definition
Bit (s)
Station Count Reported
AC 0 (VO) 1
AC 1 (VI) 2
AC 2 (BE) 3
AC 3 (BK)
4-7
Reserved
Lei Du, DoCoMo Beijing Labs
John Doe, Some Company

10. Access Category Channel Utilization IE
Month Year
doc.: IEEE yy/xxxxr0
Nov. 2006
Access Category Channel Utilization IE
Element ID
Length
Channel Utilization Bitmask
Channel Utilization List
Octets: 1
1 x (total number of non-zero bits in Channel Utilization Bitmask)
Figure X2--- AC Channel Utilization element format
Definition:
- The Channel Utilization Bitmask field indicates the AC values that have Channel Utilization specified in the following Channel Utilization List.
- The Channel Utilization List comprises a sequence of Channel Utilization fields corresponding respectively to the non-zero bits in the Channel Utilization Bitmask field.
- The Channel Utilization field is defined as the percentage of time, normalized to 255, the QAP sensed the medium was busy arising from the corresponding AC traffic.
Table 44b－ Channel Utilization Bitmask Definition
Bit (s)
CH. Utilization Reported
AC 0 (VO) 1
AC 1 (VI) 2
AC 2 (BE) 3
AC 3 (BK)
4-7
Reserved
Lei Du, DoCoMo Beijing Labs
John Doe, Some Company

11. Procedures and Usages (1)
Month Year
doc.: IEEE yy/xxxxr0
Nov. 2006
Procedures and Usages (1)
Passive scanning:
Active scanning: AP
STAx AP
STAx
Probe Request
Beacon (AC1 STA Count/Channel Utilization, AC2 STA Count/Channel Utilization…)
Probe Response (AC1 STA Count/ Channel Utilization )
STA considers the QoS-specified network load, and selects target BSS according to some selection criterion.
STA considers the QoS-specified network load, and selects target BSS according to some selection criterion.
(Re)Association Request
(Re)Association Request
(Re)Association Response
(Re)Association Response
Lei Du, DoCoMo Beijing Labs
John Doe, Some Company

12. Procedures and Usages (2)
Month Year
doc.: IEEE yy/xxxxr0
Nov. 2006
Procedures and Usages (2)
STA Statistics Request/Report:
STAx AP
STA Statistics Request
STA Statistics Report ( AC1 Station Count / Channel Utilization )
STA considers the QoS-specified network load, and selects target BSS according to some selection criterion.
(Re)Association Request
(Re)Association Response
Lei Du, DoCoMo Beijing Labs
John Doe, Some Company

13. Procedures and Usages Passive scanning: Active scanning:
Month Year
doc.: IEEE yy/xxxxr0
Nov. 2006
Procedures and Usages
Passive scanning:
Beacon contains the AC Station Count &AC Channel Utilization IEs.
Optionally, some high-prioritized AC related information, e.g. VO and VI, can be included using the bitmask field for better Beacon efficiency.
Active scanning:
Probe Response includes the AC Station Count & AC Channel Utilization IEs.
Optionally, some specific AC related information can be included using the bitmask field depending on requirement of respective STAs.
STA Statistics Request/Report:
STA Statistic Report specifies the station count & channel utilization corresponding to different ACs.
Lei Du, DoCoMo Beijing Labs
John Doe, Some Company

14. Summary and Recommendation
Month Year
doc.: IEEE yy/xxxxr0
Nov. 2006
Summary and Recommendation
The proposed AC Station Count & AC Channel Utilization IEs enables the QoS-aware load balancing by specifying the QBSS load in terms of access category.
Part of subfields corresponding to specific AC is allowed for the frame efficiency.
We recommend the AP selection strategy included in technical requirements for v.
Lei Du, DoCoMo Beijing Labs
John Doe, Some Company

15. Back up Nov. 2006 Month Year doc.: IEEE 802.11-yy/xxxxr0
Lei Du, DoCoMo Beijing Labs
John Doe, Some Company

16. Extended QoS-aware scheme – Min. number of VoIP HT
Month Year
doc.: IEEE yy/xxxxr0
Nov. 2006
Extended QoS-aware scheme – Min. number of VoIP HT
* Design Goal: Consider access priority of VoIP STAs during (re)association
* Proposal: Estimate the number of VoIP HTs by STA respectively
* Key points: Select the AP with the least VoIP hidden terminals
Selection criterion for VoIP STAs:
Min. (Tdata * (Nv-Nhv))
Tdata = Ldata / vm(SNR)
Nv-Nhv: number of hidden VoIP STAs
Nvi : number of associated VoIP STAs in APi
Nhvi: heard number of VoIP STAs when scanning APi for a time period Ts
Ts: Scanning period on each CH, no shorter than HearIntvmin till a beacon Reception
Tdata: Data transmission time
Vm(SNR): Max. supportable data rate determined by SNR of received beacon
Lei Du, DoCoMo Beijing Labs
John Doe, Some Company

17. Operation steps Nov. 2006 Assume Mixed VoIP and BE STAs as example
Month Year
doc.: IEEE yy/xxxxr0
Nov. 2006
Operation steps
Assume Mixed VoIP and BE STAs as example
1) Passive association: By sensing Beacon frame periodically
Nv1: VoIP STA Count in AP1
Nv2: VoIP STA Count in AP2
Nb1: BE STA Count in AP1
Nb2: BE STA Count in AP2
Nhv1: Num. of Heard VoIP STA in AP1
Nhv2: Num. of Heard VoIP STA in AP2
Nv1
Nb1
Nv2
Nb2
Select AP with
Min. (Nv - Nhv)*Tdata
New user
HearIntvmin
HearIntvmin
Scan f1
Scan f2
Association
Count number of heard VoIP STAs Nhv1
Count number of heard VoIP STAs Nhv2
Lei Du, DoCoMo Beijing Labs
John Doe, Some Company

18. Operation steps Nov. 2006 2) Active association Probe Response Nv1 Nb1
Month Year
doc.: IEEE yy/xxxxr0
Nov. 2006
Operation steps
AP1
Associated with AP1
AP2
Associated with AP2
MT3
MT1
MT2
2) Active association
By Probe Request / Probe Response frame exchange
Probe Response
Nv1
Nb1
Nv2
Nb2
Select AP with Min. (Nv-Nvh)*Tdata
HearIntvmin
Association
Count number of heard VoIP STAs Nhv1
Count number of heard VoIP STAs Nhv2
Lei Du, DoCoMo Beijing Labs
John Doe, Some Company

19. Simulation Scenario Nov. 2006
Month Year
doc.: IEEE yy/xxxxr0
Nov. 2006
Simulation Scenario
2 APs using non-overlapping channels, no co-channel interference
14 VoIP + 4 BE STAs, unevenly distributed in AP1 and AP2
Simulate the 10 seconds after given number of users joined sequentialy
IEEE e EDCA:
UL: - VoIP: AIFS = DIFS, CWmin/CWmax = 7/15, Retrylimit = 3,
select AP with min. hidden AC1 STA/Channel utilization
- BE: AIFS = DIFS+SlotTime, CWmin/CWmax = 31/1023, Retrylimit = 15
select AP with min. hidden total STA/Channel utilization
DL: VoIP packets are sent as long as it is buffered in AP.
No RTS/CTS, packets are dropped after staying in buffer > 200ms
Available data rate
[5.5, 11]Mbps
Rxthreshold
[-84, -81]dB
Path loss ratio 3
* Simulated algorithm:
1. Max. SNR (i.e. min.distance)
2. Min. Load (min. NMT)
3. Proposal : Min.(N-Nh)*Tdata
4. Ext. Proposal : Min.(Nv-Nhv)*Tdata
Lei Du, DoCoMo Beijing Labs
John Doe, Some Company

20. Simulation Model Nov. 2006 * Traffic Model: * Simulation Parameters:
Month Year
doc.: IEEE yy/xxxxr0
Nov. 2006
Simulation Model
* Traffic Model:
User Poisson Arrival:
* Simulation Parameters:
User1
User2
Exponential, Mean = 0.5 second
Parameters
Value
Transmission Range
100 m
Distance between APs
40m
Packet Size (VoIP/BE)
200/500 bytes
Beacon Interval
100 ms
HearIntvmin
10 ms
Max. Hear Interval
Reassociation Cycle
1s, if PLR>0.1
Simulation Time
10 s after all users join …
On: CBR, pktsize = 200 bytes, Exponential, Mean = 1 s
Off: Exponential, Mean = s
VoIP Traffic Model: … …
20ms
Pareto, A = 1.1, k=4.5kbytes,m = 2Mbytes
BE Traffic Model: … … …
Exponential, 6ms
Exponential, Mean = 1 second
Lei Du, DoCoMo Beijing Labs
John Doe, Some Company

21. Preliminary Simulation Results
Month Year
doc.: IEEE yy/xxxxr0
Nov. 2006
Preliminary Simulation Results
Number of STAs in AP1/AP2
UL VoIP Throughput (Mbits/sec)
VoIP PacketLoss Rate (PLR)
Number of STAs with PLR > 0.1
Max.SNR
3/11 (VoIP)
1/3 (BE)
0.64
0.07 4
Min.Nmt
9/5 (VoIP)
0.643
0.066 3
Proposal1
7/7 (VoIP)
0.66
0.038 2
Proposal2
0.68
0.03 1
Lei Du, DoCoMo Beijing Labs
John Doe, Some Company

22. Month Year
doc.: IEEE yy/xxxxr0
Nov. 2006
Conclusion
The use of AC STA Count/AC CH Utilization enables different AP selection strategy for different ACs.
The access priority for different ACs can be considered using the AC STA Count/AC CH Utilization instead of the general total values.
Lei Du, DoCoMo Beijing Labs
John Doe, Some Company