15
doc.: IEEE 802.11-13/1153r0 Submission September 2013 Laurent Cariou (Orange) Slide 1 Simulation scenario proposal Date: 2013-17-09 N am e C om pany A ddress Phone em ail LaurentCariou O range 4 rue du closCourtel 35512 Cesson Sévigné France +33 299124350 [email protected] Thom asD erham O range 9F K eio Shinjuku O iw ake Bldg. Shinjuku 3-1-13, Tokyo, Japan +81 3 5312 8563 thomas.derham @ orange.com Jean-Pierre Le Rouzic O range 4 rue du ClosCourtel 35512 Cesson Sevigne France +33 299124893 [email protected] Authors:

Doc.: IEEE 802.11-13/1153r0 Submission September 2013 Laurent Cariou (Orange)Slide 1 Simulation scenario proposal Date: 2013-17-09 Authors:

Embed Size (px)

Citation preview

Page 1: Doc.: IEEE 802.11-13/1153r0 Submission September 2013 Laurent Cariou (Orange)Slide 1 Simulation scenario proposal Date: 2013-17-09 Authors:

doc.: IEEE 802.11-13/1153r0

Submission

September 2013

Laurent Cariou (Orange)Slide 1

Simulation scenario proposal

Date: 2013-17-09

Name Company Address Phone email Laurent Cariou Orange 4 rue du clos Courtel

35512 Cesson Sévigné France

+33 299124350 [email protected]

Thomas Derham Orange 9F Keio Shinjuku Oiwake Bldg. Shinjuku 3-1-13, Tokyo, Japan

+81 3 5312 8563

[email protected]

Jean-Pierre Le Rouzic Orange 4 rue du Clos Courtel 35512 Cesson Sevigne

France

+33 299124893 [email protected]

Authors:

Page 2: Doc.: IEEE 802.11-13/1153r0 Submission September 2013 Laurent Cariou (Orange)Slide 1 Simulation scenario proposal Date: 2013-17-09 Authors:

doc.: IEEE 802.11-13/1153r0

Submission

Context

• Simulation scenarios definition is a key step to clearly scope the work in HEW.

• We need to ensure that they address the real issues (scenario / interference / traffic model)

– while being simple enough

– and being “open enough” to not preclude parts of the potential solution space

• We take the assumptions that HEW will have to ensure improvements in « all main scenarios »

– we can therefore spread the difference interference issues over the different scenarios

• In this presentation, we review the main simulation scenarios and– allocate specific interference condition problems to specific scenario(s)

– define some options in the scenarios to capture all elements that can impact the performance of specific solutions

Slide 2 Laurent Cariou (Orange)

September 2013

Page 3: Doc.: IEEE 802.11-13/1153r0 Submission September 2013 Laurent Cariou (Orange)Slide 1 Simulation scenario proposal Date: 2013-17-09 Authors:

doc.: IEEE 802.11-13/1153r0

Submission

Define simulation scenarios to capture all issues and to enable solutions evaluation

1 Need to make sure that all interference conditions are captured and spread over the different scenarios

2 Need to enable PHY/MAC simulation parameters to be tuned, not to restrict the solution space (e.g. TxPower, channel selection…)

3 Need to include different options in the simulation scenarios so that, for the evaluation of specific schemes, all traffic or devices, that can have an impact on the performance, are taken into account

– these options should be assigned to specific family of solutions as part of the evaluation methodology/selection procedure

Slide 3 Laurent Cariou (Orange)

September 2013

Page 4: Doc.: IEEE 802.11-13/1153r0 Submission September 2013 Laurent Cariou (Orange)Slide 1 Simulation scenario proposal Date: 2013-17-09 Authors:

doc.: IEEE 802.11-13/1153r0

Submission

1 Mapping of interference types on the main simulation scenarios

• We review the types of interference defined in the usage model document, and evaluate if they are captured in the current simulation scenarios– We base our analysis on document 1000r0, and mapped previous propositions,

especially 722-23r0

• The parameters that currently define the interference scenario are mostly the topology (AP/STA placements, channel conditions…) and whether the network is managed or not

– By managed, it is meant that multiple APs are sharing the same management entity

– But more importantly for simulation scenario definition: a managed network is planned, meaning that the AP locations follow a relatively regular grid and that specific frequency reuse schemes can be applied.

Slide 4 Laurent Cariou (Orange)

September 2013

Page 5: Doc.: IEEE 802.11-13/1153r0 Submission September 2013 Laurent Cariou (Orange)Slide 1 Simulation scenario proposal Date: 2013-17-09 Authors:

doc.: IEEE 802.11-13/1153r0

Submission

1 Mapping of interference types on the main simulation scenarios

• 1 Residential:– fully unmanaged (unplanned) networks, potentially overlaid by some P2P links

– topology: High density of BSS, low to medium STA density, indoor

• 2a Enterprise:– Managed (planned) ESS overlaid by many unmanaged P2P links

– topology: high density of BSSs, high density of STAs, indoor

• 2b Dense indoor hotspot:– Managed ESS (currently no overlaid interference proposed)

– topology: High density of BSSs, high density of STAs, indoor

• 3a Outdoor large deployment:– managed ESS (currently no overlaid interference proposed)

– topology: relatively low density of BSSs, high density of STAs, outdoor

Slide 5 Laurent Cariou (Orange)

September 2013

Page 6: Doc.: IEEE 802.11-13/1153r0 Submission September 2013 Laurent Cariou (Orange)Slide 1 Simulation scenario proposal Date: 2013-17-09 Authors:

doc.: IEEE 802.11-13/1153r0

Submission

1 Mapping of interference types on the main simulation scenarios

  Interference in unplanned deployments

Interference between OBSSs belonging to the same planned ESS (cell-edge with low SNR)

Interference between OBSSs belonging to the same planned ESS (cell-edge with high SNR)

Interference from P2P unmanaged OBSS on a planned ESS

interference from multiple overlapping planned ESSs

Residential  captured (723)        

Enterprise      captured (723)  captured (723)

 not captured

Dense Hotspot      captured (723)  not yet captured

 not captured

Outdoor large deployment

   captured (723)    not captured  not yet captured

Slide 6 Laurent Cariou (Orange)

September 2013

Page 7: Doc.: IEEE 802.11-13/1153r0 Submission September 2013 Laurent Cariou (Orange)Slide 1 Simulation scenario proposal Date: 2013-17-09 Authors:

doc.: IEEE 802.11-13/1153r0

Submission

1 Mapping of interference types on the main simulation scenarios: Focus on dense hotspot 2b

• As the dense hotspot scenario is managed, we consider a frequency reuse deployment

• ICD is low (10-20m)

• Interference between BSSs belonging to the same planned ESS is important and well captured

BSSBSS

ClusterCluster

BSSBSS

BSSBSS

BSSBSSBSSBSS

BSSBSSBSSBSS

BSSBSS

simulation of all channelswith frequency reuse 3

Simplified simulation: One channel only, assuming frequency reuse 3

CCA range

STAs are concentrated close to the AP:good SNR at cell-edge

Slide 7 Laurent Cariou (Orange)

September 2013

Page 8: Doc.: IEEE 802.11-13/1153r0 Submission September 2013 Laurent Cariou (Orange)Slide 1 Simulation scenario proposal Date: 2013-17-09 Authors:

doc.: IEEE 802.11-13/1153r0

Submission

• We should add interfering networks to the dense hotspot scenario – overlay of BSS short-range links (for tethering devices)

– randomly distributed on the whole simulation zone• some of these BSS can be idle and

only transmit management frames

– Note that there are similarities with Enterprise 2a

BSSBSS

BSSBSS

BSSBSSBSSBSS

BSSBSSBSSBSS

BSSBSS

1 Mapping of interference types on the main simulation scenarios: Focus on dense hotspot 2b

Slide 8 Laurent Cariou (Orange)

September 2013

Page 9: Doc.: IEEE 802.11-13/1153r0 Submission September 2013 Laurent Cariou (Orange)Slide 1 Simulation scenario proposal Date: 2013-17-09 Authors:

doc.: IEEE 802.11-13/1153r0

Submission

• As the outdoor large hotspot scenario is managed, we consider a frequency reuse deployment as well

• ICD is large: 100-150m

• With this frequency reuse, interference between BSSs from the managed network is very low

BSSBSS

ClusterCluster

BSSBSS

BSSBSS

BSSBSSBSSBSS

BSSBSSBSSBSS

BSSBSS

All channels with frequency reuse 3

One channel with frequency reuse 3

CCA range

STAs are distributed on the whole AP coverage:low SNR at cell-edge

1 Mapping of interference types on the main simulation scenarios: Focus on large outdoor 3a

Slide 9 Laurent Cariou (Orange)

September 2013

Page 10: Doc.: IEEE 802.11-13/1153r0 Submission September 2013 Laurent Cariou (Orange)Slide 1 Simulation scenario proposal Date: 2013-17-09 Authors:

doc.: IEEE 802.11-13/1153r0

Submission

• Potential interfering network on outdoor hotspot – overlay of other operators will be very common in such scenarios and is the

most different interference from other scenarios

– in the case of an overlay of 3 operators planned deployments, each of which with freq reuse 3, the scenario becomes a simple theoretical frequency reuse 1 deployment with 100-150m ICD between neighboring OBSSs

• with an overlap between neighboring cell around cell-edge (low SNR)

• with neighboring cells not sharing the same management entity

• with constraint on associations: STAs connect only to one operator APs

BSSBSS

BSSBSS

BSSBSS

BSSBSS

BSSBSS

BSSBSS

1 operator 2 operators 3 operators frequency reuse 1deployment

1 Mapping of interference types on the main simulation scenarios: Focus on large outdoor 3a

Slide 10 Laurent Cariou (Orange)

September 2013

Page 11: Doc.: IEEE 802.11-13/1153r0 Submission September 2013 Laurent Cariou (Orange)Slide 1 Simulation scenario proposal Date: 2013-17-09 Authors:

doc.: IEEE 802.11-13/1153r0

Submission

2 Allow parameters tuning to evaluate technical propositions

• We all agree that on each scenarios, we should define a default value for all parameters– to define a baseline to compare with– Once we agree on the template (1001rx) for simulation scenario, we’ll make a

proposal based on previous slides for this

• Now some of those parameters are key to the scenario definition and shouldn’t be tuned

– AP/STA positioning, channel models…– channel and potentially bandwidth allocation for planned deployments (dense hotspot and

outdoor large deployment)

• Depending on the proposed technical solutions, some parameters need to be tunable, as their adaptation is part of the solution space– Tx Power, CCA level, CSMA parameters, bandwidth, channels…

Slide 11 Laurent Cariou (Orange)

September 2013

Page 12: Doc.: IEEE 802.11-13/1153r0 Submission September 2013 Laurent Cariou (Orange)Slide 1 Simulation scenario proposal Date: 2013-17-09 Authors:

doc.: IEEE 802.11-13/1153r0

Submission

3 Define options in the simulation scenarios: proportion of HEW and legacy devices

• For all the technical propositions, especially for OBSS, we need to model all elements that could jeopardize the performance of the scheme (to ensure it will keep its efficiency in real deployments)

• If the presence of « legacy devices » is susceptible to have such degrading impact :

– It will be important to define a proportion (TBD%) of devices in the scenario that won’t implement the proposed scheme and would keep the baseline default parameters

• STAs connected to the planned network• APs and STAs part of the interfering network

– whether this is mandatory or not should be determined by the group on a per solution basis as part of the evaluation methodology/selection procedure

• This « optional/mandatory » mix of devices should be defined in the simulation scenario

Slide 12 Laurent Cariou (Orange)

September 2013

Page 13: Doc.: IEEE 802.11-13/1153r0 Submission September 2013 Laurent Cariou (Orange)Slide 1 Simulation scenario proposal Date: 2013-17-09 Authors:

doc.: IEEE 802.11-13/1153r0

Submission

3 Define options in the simulation scenarios: proportion of HEW and legacy devices

Example with Transmit power control and CCA control

• For example, power control solutions are showing promising theoretical gains in dense deployments

• However, it is well understood that the use of different power or CCA levels by devices leads to throughput starvation of certain nodes due to the introduction of asymmetric links– some power control solutions can therefore become inefficient

• if applied only in the AP but not on STA side• in the presence of overlapping BSSs (P2P links) which don’t apply TPC, or apply it differently• if only part of the traffic is applying TPC (APs could apply TPC for data transmission but not

for management frames)

• If we want to make sure that such a proposed solution is efficient in real deployments, we would need to capture these degrading elements in the simulation scenarios :– proportion of devices (connected STAs and AP/STAs from the interfering networks) that

don’t implement TPC and use « default » TxPower– idle AP/STAs that don’t apply TPC to their management frames transmission

Slide 13 Laurent Cariou (Orange)

September 2013

Page 14: Doc.: IEEE 802.11-13/1153r0 Submission September 2013 Laurent Cariou (Orange)Slide 1 Simulation scenario proposal Date: 2013-17-09 Authors:

doc.: IEEE 802.11-13/1153r0

Submission

Conclusion

• We propose a design for dense hotspot 2b and outdoor large BSS 3a scenario– we’ll use template 1001rx (if/when agreed by the group) to make a proposition for default

parameters for these scenarios, based on the current presentation

• Enabling parameter tuning by HEW devices in these scenarios should of course be enabled with some restrictions depending on the scenarios

• Options should be defined in the scenarios in order to capture all traffic or types of devices that can have an impact on specific solutions

– Example for the mix of HEW and legacy devices for TPC solutions

• In the evaluation methodology/selection procedure, each family of proposed solutions should be linked to specific simulation scenario(s) and their option(s)

– to ensure a good quality of the evaluation

– to be able to compare competing solutions

Slide 14 Laurent Cariou (Orange)

September 2013

Page 15: Doc.: IEEE 802.11-13/1153r0 Submission September 2013 Laurent Cariou (Orange)Slide 1 Simulation scenario proposal Date: 2013-17-09 Authors:

doc.: IEEE 802.11-13/1153r0

Submission

References

11-13/1000, “HEW simulation scenarios”, Simone Merlin (Qualcomm)

11-13/1001, “HEW simulation scenarios document template”, Simone Merlin (Qualcomm)

11-13/0722, “HEW Evaluation Methodology”, Minyoung Park (Intel)

11-13/0723, “HEW SG evaluation methodology overview” Minyoung Park (Intel)

11-13/1051, “HEW evaluation methodology” Ron Porat (Broadcom)

September 2013

Slide 15 Laurent Cariou (Orange)