cjk test-bed study performance measurements of rtp/rtcp based services
DESCRIPTION
CJK test-bed study Performance measurements of RTP/RTCP based services. 15th NGN-WG 200 9 . 4.8-10 Test-bed Ad-hoc Group Hideaki YAMADA KDDI (KDDI R&D Labs.) [email protected] TEL: +81 49 278 7356, FAX: +81 49 278 7510. Topics. Background Preparations - Study phases (Study scenarios) - PowerPoint PPT PresentationTRANSCRIPT
CJKTest-bed
CJKTest-bed
CJK test-bed studyPerformance measurements of RTP/RTCP based services
15th NGN-WG 2009.4.8-10Test-bed Ad-hoc Group
Hideaki YAMADA KDDI (KDDI R&D Labs.)
[email protected]: +81 49 278 7356, FAX: +81 49 278 7510
2
Topics
Background Preparations- Study phases (Study scenarios)- Measurement points
Performance measurementStandardization-Y.2173 :Management of Performance Measurement(MPM)-RACF and MPM communications
3
Background
1) Estimation of QoE from QoS
Effective utilization of network resources considering the QoE
Service platform for MoIP based on the end-to-end QoE in FMBC (Fixed Mobile & Broadcast Convergence) all-IP networks
2) Dynamic session controls of MoIP services in FMBC all-IP networks
4
CJK NGN Test-bed
Phase I: 2006 3rd/4th Quarter (Completed)- Network Connectivity- Scenario 2 and 4 only (2 CS and 2 domains)
Phase II: 2007 1st/2nd Quarter (Completed)- Scenario 1, 3 (simpler version of 2 and 4, single domain) and 5
Phase IV: 2008 3rd ~ 2009 2nd Quarter- Performance Evaluation of RTP/RTCP-based MoIP and IPTV services- Performance Monitoring Scenarios 6-3 and 6-4- Initial RACF Interoperability testing : Scenario 7- Testing of RACF and RTP/RTCP-based MPM Interactions- IPTV Interoperability testing
Preparations - Study phases (Study scenarios) 1
5
China (CCSA) Korea (TTA) Japan (TTC)
IP Network C-K: R&D network C-J, K-J: Commercial networks(or R&D network)
CPE for Voice over IP
CPE used in Phase I&II are applicable • Software based CPE supporting the RTCP extensions such as SR/RR/XR/HR
for Video over IP
TBD TBD • Hardware based CPE• Support RTCP SR/RR/XR video
Service Platform Call servers used in Phase I&II 1st Stage: Call server emulator (planning)2nd Stage: IMS
MPM MPM (Software/Hardware)1st Stage: PC based MPM provided by KDDI R&D labs.2nd Stage: ATCA based MPM (based on the SBC by OKI Electric)
RACF TBD In preparation TBD
Preparations - Study phases (Study scenarios) 2
6
Japan (TTC)
Korea (TTA)
MPM (Router with mirror port and RTP translator)
China (CCSA)
CPE (PC with RTP/RTCP
terminal software)
CPE (RTP/RTCP terminal hardware)
MPM
SIP Server SIP Server
SIP Server
Scenario 6-3
Scenario 6-4, 7
CPE (PC with RTP/RTCP
terminal software)
CPE (RTP/RTCP terminal hardware)
CPE (PC with RTP/RTCP
terminal software)
CPE (RTP/RTCP terminal hardware)
MPM
Scenario 6-1
Scenario 6-2
RACF
RACF
RACF
Preparations - measurement points
7
Reference : ETSI TR 101 329-7 V2.1.1, February 2002.
The E-model rating called R-value (ITU-T G.107) against the intrinsic delay for a single MPLS network.
Performance measurement 1
8
A result of the PESQ(Perceptual Evaluation of Speech Quality) score versus the Jsd value which is a statistical factor of VoIP packets called standard deviation of jitters. The codec type is the ITU-T G.711 with no packet loss.
Figure 2: PESQ scores versus Jsds (standard deviation of packets’ jitters)
Performance measurement 2
Better quality
9
Transport stratum
Service stratum
ControlMediaManagement
ANI
Transport Control Functions
Resource and Admission
Control Functions
Network AttachmentControl Functions
Network AttachmentControl Functions
NNIUNI
Application Support Functions & Service Support Functions
Applications
Transport Functions
End-UserFunctions
OtherNetworks
Service ControlFunctions
Service UserProfiles
Service UserProfiles
Transport User Profiles
Transport User Profiles
MPM
MPMApplications
MPMIn
OtherNGNs
Overall architecture for NGN performance management
Standardization 1 Y.2173: Management of Performance Measurement (MPM)
10
Standardization 2 RACF and MPM communications
Figure1 Performance notification flow of MPMs
Figure 2 ITU-T Y.2111(RACF IF(Rm, Ri, Rh)
11
6.6 Resource control and scenarios for RACF and MPM communications• It is recommended to support a communication between RACF and MPM under the
following situations:• In the NGN transport network, abnormal situations where such as congestion, packet loss,
and delay can occur despite of bandwidth resource availability due to various reasons. The causes can be physical such as link failure and system down or due to logical network failure (.e.g., routing loop), logical like routing loop. The physical failure can be reported to the RACF so that it can take appropriate actions during the admission decision process. However, the logical failure can’t be reported to RACF under current architecture. This will results in the inefficient or even wrong admission decision because RACF depends mainly on the topology and bandwidth resource information for its decision making. Thus, some additional performance information besides bandwidth such as delay, delay variation, and packet loss will help to make more accurate admission decision when such logical failure situation occurs..
• In normal situation, in addition to bandwidth, network policy rules, and subscription information, performance information such as delay, delay variation, and packet loss may still be needed when RACF makes admission decision.
• Upon receipt of an explicit request from SCF including an indication of specific performance monitoring, RACF can communicate with MPM to perform the request and receives status reports of the requested performance monitoring.
Standardization 3 RACF and MPM communications (Rec. Y.2111 Revision 2 (Version 0.1.0))
12
Thank you for your kind attentions.
Q&A
This work is partly supported by the National institute of Information and Communications Technology (NICT).