doc.: ieee 802.11-13/1406r0 submission nov 2013 huai-rong shao, et al. (samsung)slide 1 traffic...
TRANSCRIPT
doc.: IEEE 802.11-13/1406r0
Submission
Nov 2013
Huai-Rong Shao, et al. (Samsung)Slide 1
Traffic Modeling for HEW Simulation
Date: 2013-11-12
Name Affiliations Address Phone email
Huai-Rong Shao Samsung [email protected]
Chiu Ngo Samsung [email protected]
SangHyun Chang Samsung [email protected]
Hyunjeong Kang Samsung [email protected]
Guoqing Li Intel [email protected]
Saishankar Nandagopalan
Adeptence Inc [email protected]
Yonggang Fang ZTETX [email protected]
Bo Sun ZTE [email protected]
Kaiying Lv ZTE [email protected]
Chaochun Wang MediaTek [email protected]
Authors:
doc.: IEEE 802.11-13/1406r0
Submission Slide 2
Motivation
• HEW group tries to improve performance such as efficiency and QOE for the real world scenario
• Traffic modeling becomes very important to simulate the real use scenarios
Nov 2013
Huai-Rong Shao, et.al. (Samsung)
doc.: IEEE 802.11-13/1406r0
Submission Huai-Rong Shao, et.al. (Samsung)
Traffic Modeling Framework • Two levels for simulation:
• Level one– Application Traffic Modeling [2,3]
• Specify the traffic characteristics of an application• E.g., Local file transfer, HTTP web browsing, Lightly Compressed video, Online
gaming, etc.
• Level two– STA Application Profiles [4]
– Mixing of applications at STAs – E.g., STA 1 has a profile of video streaming + web browsing, and STA 2 has a
profile of video conferencing + ftp
• Profile Configuration (Addressed by this presentation) – Pattern of the application events, i.e., when to start the applications in the
simulation and how long for each application– E.g., for a STA, video streaming application starts at time 0 and web browsing
starts at time 8
Nov 2013
Slide 3
doc.: IEEE 802.11-13/1406r0
Submission Huai-Rong Shao, et.al. (Samsung)
Example 1: TCP Traffic Configuration
• TCP configuration for file transfer [2]
• Widely used by different simulators
Nov 2013
Slide 4
TCP Configuration Parameters in SimulatorsMSS Ethernet (1500 bytes)
Receive Buffer (Bytes) 65535
Receive Buffer Adjustment None
Delayed ACK Mechanism Segment/Clock based
Max. ACK Delay (Sec) 0.05
Slow Start Initial Count (MSS) 1
Fast Retransmit Enabled
Duplicate ACK Threshold 3
Fast Recovery Reno
Window Scaling Enabled
Selective ACK (SACK) Disabled
ECN Capability Disabled
Segment Send Threshold Byte Boundary
Active Connection Threshold Unlimited
Karn’s Algorithm Enabled
Nagle Algorithm Disabled
Initial Sequence Number Auto Complete
Initial RTO (sec) 3.0
Min RTO (sec) 1.0
Max RTO (sec) 64
RTT Gain 0.125
Deviation Gain 0.25
RTT Deviation Coefficient 4.0
Timer Granularity 0.25
doc.: IEEE 802.11-13/1406r0
Submission
Nov 2013
Huai-Rong Shao, et.al (Samsung)Slide 5
Example 2: HTTP Traffic Modeling
• The parameters of HTTP file transfer in [2]:– SM: Size of main object in page– Nd: Number of embedded objects
in a page– SE: Size of an embedded object in
page– Dpc: Reading time– Tp: Parsing time for the main page
• The table on the right highlights the components and their distributions– Approximate mean data rate of
(10KB+48KB)/31 = 12.31 kbps
Component Distribution Parameters PDF
Main object size (SM)
Truncated Lognormal
Mean = 10710 bytes SD = 25032 bytes Min = 100 bytes Max = 2 Mbytes (before truncation)
21 lnexp , 0
22 2
xf xx
x
37.8,37.1
if x>max or x<min, discard and generate a new value for x
Embedded object size (SE)
Truncated Lognormal
Mean = 7758 bytes SD = 126168 bytes Min = 50 bytes Max = 2 Mbytes (before truncation)
21 lnexp , 0
22 2
xf xx
x
17.6,36.2
f x>max or x<min, discard and generate a new value for x
Number of embedded objects per page (Nd)
Truncated Pareto
Mean = 5.64 Max. = 53 (before truncation)
,1
,
k k x mf xx
kx mf x
m
1.1, 2, 55k m
Subtract k from the generated random value to obtain Nd if x>max, discard and regenerate a new value for x
Reading time (Dpc)
Exponential Mean = 30 sec
, 0x
xf ex
= 0.033
Parsing time (Tp)
Exponential Mean = 0.13 sec
0,
xex
f x
69.7
doc.: IEEE 802.11-13/1406r0
Submission
Nov 2013
Huai-Rong Shao, et.al (Samsung)Slide 6
Modeling for Various Video Traffic Types
• [3] provides models for – Wireless Display: lightly compressed video– Buffered video streaming– Video conferencing
• Details discussed during this week
doc.: IEEE 802.11-13/1406r0
Submission Huai-Rong Shao, et.al. (Samsung)
Application Event Modeling
• How to simulate the event timing of individual application traffic for all STAs to simulate real world scenario– When to generate which traffic types– Duration of the applications, etc.
• Different use scenarios may need different application event models – Classroom, Home, Railway station, Stadium, etc.– Some may completely random and some may have peak patterns
Nov 2013
Slide 7
doc.: IEEE 802.11-13/1406r0
Submission Huai-Rong Shao, et.al. (Samsung)
Application Event Patterns
• Most popular pattern is Random pattern– A Random pattern occurs where there are many
users, each generating a little bit of traffic and requesting network access randomly
– Currently Poisson model is widely used• VoIP calls from Cisco measurement [5]• User accesses to Web Pages [6]
• The second pattern is Peaked pattern– A peaked pattern has big spikes from the mean
• Some use scenarios like training arrival, class starting may go to this pattern
– Hyper-exponential model can be used• VoIP calls from Cisco measurement [5]
Nov 2013
Slide 8
Time
Application requests
Application requests
Time
Time
doc.: IEEE 802.11-13/1406r0
Submission
Summary• HEW simulation needs both application traffic models
and application profiling/configuration models– We have various applications traffic models [2,3] proposed
• More traffic types may be added in the future– Application profiling/configuration
• Based on [4], add application event models
• Propose to add application event models to Simulation Scenario document (802.11-13/1001)
• Poisson model and Hyper-exponential model
Nov 2013
Huai-Rong Shao, et.al (Samsung)Slide 9
doc.: IEEE 802.11-13/1406r0
Submission
Reference
[1] “HEW SG Simulation Scenarios”, IEEE 802.11-13/1001r4, Simone Merlin, etc.
[2] “Internet Traffic Modeling”, IEEE 802.11-09/1317r1, Sai Nandagopalan, etc.
[3] “Video Traffic Modeling”, IEEE 802.11-13/1305r0, Guoqing Li, etc.
[4] “Simplification of HEW Traffic Model Simulations”, IEEE 802.11-13/0847r1, William Carney, etc.
[5] “Traffic Analysis for Voice over IP”, http://www.cisco.com/en/US/docs/ios/solutions_docs/voip_solutions/TA_ISD.html#wp1030552
[6] “A Poisson Model for User Accesses to Web Pages”, Computer and Information Sciences -- ISCIS 2003, Sule Gunduz, etc.
Slide 10
Nov 2013
Huai-Rong Shao, et.al. (Samsung)
doc.: IEEE 802.11-13/1406r0
Submission Huai-Rong Shao, et.al. (Samsung)
Thank you!
Nov 2013
Slide 11