Ansi/tia-921-b network Model for Evaluating Multimedia Transmission Performance Over the Internet Protocol



Download 1.57 Mb.
Page4/17
Date31.01.2017
Size1.57 Mb.
#13070
1   2   3   4   5   6   7   8   9   ...   17

4.4Interfering Stream Files


The network traffic files used by the simulator are all of a standard packet capture format commonly known as pcap, as defined in file pcap.h (version 2.4) of libpcap. See http://wiki.wireshark.org/Development/LibpcapFileFormat . Usually the files have been captured at the endpoints of the network at client devices. These files are used as interfering traffic input to the simulation model.
The following types of streams are included in the standard test cases:

  • IPTV: SD and HD, CBR and VBR

  • VoIP

  • FoIP (fax over IP)

  • Peer-to-peer file transfer

  • HTTP

  • OTT TCP video: SD and HD

These files have been anonymously captured by a variety of researchers and from a variety of sources as real examples of representative types of packet traffic. In general, the files are a reflection of the traffic patterns observed and have had the payloads stripped to protect the privacy and rights of the actual content as well as to keep the overall file size small. Note that even though the traffic source files omit the payloads, size information is maintained so that the full traffic is modeled by the simulation. Pcap files with payloads can be used as test streams or interfering streams, since the simulator does pass payload data.
In addition to payload stripping, the pcap files have had a number of processing steps:

  • Flow separation: The captured files are split into upstream (toward the core) and downstream (toward the client endpoint) files for simulation.

  • Bandwidth analysis: The files are analyzed as to their overall and average bandwidth over finite time periods. As reference files, the microscopic and macroscopic bandwidth characteristics must be understood so that they can be applied reasonably to the various simulation test cases. In particular, the bandwidth characteristics of the files are highly dependent on factors such as the overall bandwidth capabilities and rate limitations of the network path, and the behavior and rate adaptation of higher layer network protocols.

  • Smoothing: At a microscopic level, bandwidth analysis of some pcap files reveals that there are microbursts. A potential problem with these microbursts is that the instantaneous bandwidth requirement can overwhelm link or memory capacities of the scenario under test. This can be alleviated either through external smoothing of the file or by the rate shaping parameters in the simulation configuration file.

  • Bandwidth scaling: The standard pcap files are specific captures intended to be representative of network usage. Since they are generic in nature, playback into the simulator is made by bandwidth scaling. This is accomplished by temporal dilation: playback of packets by a constant scale factor, faster or slower, achieves the desired network usage.

6.4.1.1.1.3 describes the standard pcap files in detail, and includes the files as an electronic attachment. The average bandwidth of one example (using a five second non-overlapping window) is plotted in Figure .
c:\documents and settings\edschulz\local settings\temporary internet files\content.word\clipboard02.png

Figure : Sample Interferer – OTT2 Downstream Flow


4.5Simulation Inputs


Table lists the inputs to the simulation. The core-to-LAN network topology is assumed. One access technology (DSL or GPON) is selected. There are parameters for each network element and for each pcap file used as an interfering traffic stream.
Traffic flows are assigned a priority that is essential in practical access networks to ensure that higher priority traffic from managed services is carried in preference to lower priority traffic. The simulator has 7 priority levels and no per-class bandwidth reservation. The simulation test cases are set as follows:

  • 1: (highest priority) Managed voice traffic

  • 2: Managed IPTV-type video traffic

  • 3 - 6: Unused

  • 7: (lowest priority) Best effort, OTT video, VoIP/FoIP, peer-to-peer


Table : Network Model Simulation Input Parameters

Network Element

Impairment / Interferer

Parameter Range

Core

 

 

 

Number of Switches

3 to 15

 

Total Core Link Delay (ms)

10 to 250

 

 




Access (pick one technology)




GPON

Access Rate Down (Mbit/s)

5 to 50

 

Access Rate Up (Mbit/s)

2 to 35

 

Residual BER

10-12 to 10-9

 

Delay (ms)

1

 

 




DSL

Access Rate Down (Mbit/s)

3 to 33

 

Access Rate Up (Mbit/s)

1 to 3

 

Residual BER

10-8 to 10-6

 

Delay (ms)

1










 

Managed Bandwidth

 

 

IPTV HD Stream 1 - CBR (qty)

 0 to 1 (10 core-only)

 

Downstream Rate (Mbit/s)

8

 

IPTV HD Stream 2 - VBR (qty)

 0 to 1 (10 core-only)

 

Downstream Rate (Mbit/s)

8

 

IPTV SD Stream 1 - CBR (qty)

 0 to 1 (10 core-only)

 

Downstream Rate (Mbit/s)

2

 

IPTV SD Stream 2 - VBR (qty)

 0 to 1 (10 core-only)

 

Downstream Rate (Mbit/s)

2

 

VoIP/Fax (qty)

 0 to 1 (10 core-only)

 

Rate down/up (Mbit/s)

0.064 / 0.064

 

 

 

 

Residual Bandwidth

 

 

Peer-to-peer Rate Down / Up

1.89 / 0.768

 

HTTP Rate Down / Up

1.7 / 0.007

 

OTT1 Rate Down / Up

7 / 0.128

 

OTT2 Rate Down / Up

1 / 0.025

 

VoIP/FoIP Rate Down / Up

0.064 / 0.064





Download 1.57 Mb.

Share with your friends:
1   2   3   4   5   6   7   8   9   ...   17




The database is protected by copyright ©ininet.org 2024
send message

    Main page