網路電話品質量測技術
VoIP Quality Measurement
- Course webpage: http://course.ipv6.club.tw/Measurement/
- Time: Thursday 08:10-11:00
- Place: TC-115
- Textbook:
- Reference Book: "Carrier Grade Voice over IP" by Daniel Collins,
published by McGraw-Hill.
- Students enroll in this class
- TA:
課程評量方式
- 作業(20%)
- 課堂報告(40%)
- 期中考(20%)
- 期末考(20%)
Topics
- Performance Measurement on Internet
- VoIP Quality
- VoIP Quality Measurement
- QoS
- DiffServ
- RSVP
- MPLS
- Stress Test
- NAT issue
- SIPp
- RTP streaming
- IPv4/IPv6 Translation
Homework
Presentation
- SIPp
- 10/15 SIPp (Junn-keh Yeh)
- 10/22 SIPp XML (Kai-Jia
Chang)
- Cacti
- 10/8 BWCTL
- iPerf (Wei-Li Lai)
- http://www.internet2.edu/performance/bwctl/index.html
- http://iperf.sourceforge.net/
- BWCTL (Shin-Fu Huang)
- 11/19 Audio I/O (Wei-Shin)
- 11/26
ITU-T P.800 (Solomon)
- 12/3
PESQ Wen-Jen
- 12/17
PJSUA (Ni-Ya)
- 1/7 PJSUA (Li-Wen)
- 1/7 oRTP (Wei-Lin)
- 1/14 perfSONAR
perfSONAR (施創宏)
呂冠達
Tools
- NIST Net - A
Linux-based Network Emulation Tool (It can emulate packet loss rate.)
- Softpedia Network Tools
-
VoIP TroubleShooter
- TestYourVoIP.com
This site will establish a SIP session with your PC and
calculate
the MOS value according to the E-Model (ITU-T G.107).
- Google
Suggest
- http://www.internet2.edu/performance/pS-PS/
- http://e2epi.internet2.edu/ndt/
- CAIDA -
Performance Measruement Tools
- TWAREN Measurement
- SpeedTest.net
- Google M-Lab
(PDF)
- SNMP - MRTG
- NLANR - National Laboratory for
Applied Network Research
- Iperf - Measure bandwidth availabity using a client and server
model.
- Pathchar - determine per-hop bandwidth, propagation delay, queue time and
drop rate
-
Information on creating your own AMP network
- Internet End-to-end
Performance Monitoring (IEPM) in Stanford
- Reverse Traceroute Server
- PingER (Ping End-to-end Reporting)
Codec
- [Talk 2] OWAMP
-
IETF IP Performance Metrics (IPPM) working group
- RFC 2679 - A
One-way Delay Metric for IPPM
- This document describes some common errors and uncertainly
which you must take into account when performing measurements.
However, it is not directly related to OWAMP protocol.
- In my opinion, we do not need to "inject" traffic into the
network to perform these measurements. We can just collect the
results from RTCP.
- RFC 3763 - One-way Active Measurement Protocol (OWAMP)
Requirements
- There should exist two
protocols: one for conducting the actual measurement session and
another for session setup/teardown/confirmation/retrieval. These
protocols are referred to as OWAMP-Test and OWAMP-Control,
respectively.
Comments: This is similar to RTP and SIP!
- If OWAMP-Test packets are easy to identify
(e.g., they all come to a well-known port number), an intermediate
party might place OWAMP-Test traffic into a priority queue at a
congested link thus ensuring that the results of the measurement
appear better than what would be experienced by other traffic.
- RFC 4656 -
A One-way Active Measurement Protocol (OWAMP)
- TCP well-known port 861 is allocated to OWAMP Control.
- OWAMP test traffic is simply a stream of
UDP packets from and to negotiated port numbers.
If you don't want that the network provider can detect OWAMP traffic
easily, you may apply encryption on OWAMP packets.
- The goal of encrypted mode is: to make it hard for
a
party in the middle of the network to make results look "better" than
they should be.
- because many Internet paths include segments that transport
IP over ATM, delay and loss measurements can include the effects of
ATM segmentation and reassembly (SAR). Consequently, OWAMP has
been designed to allow for small test packets that would fit inside
the payload of a single ATM cell (this is only achieved in
unauthenticated mode).
- OWAMP-Control
communications are encrypted with the AES Session-key (using
CBC mode) and authenticated with HMAC Session-key
- SID is always chosen by the receiver, by concatenation of the
4-octet IPv4 IP
number belonging to the generating machine, an 8-octet timestamp, and
a 4-octet random value.
- Non-zero Accept values indicate a failure of some sort. Zero
values indicate normal completion.
- Each packet record is 25 octets and includes 4 octets of sequence
number, 8 octets of send timestamp, 2 octets of send timestamp error
estimate, 8 octets of receive timestamp, 2 octets of receive
timestamp error estimate, and 1 octet of Time To Live (TTL), or Hop
Limit in IPv6
- in encrypted mode both
the sequence number and the timestamp are protected to provide
maximum data confidentiality and integrity protection, whereas in
authenticated mode the sequence number is protected while the
timestamp is sent in clear text.
- Sending the timestamp in clear text
in authenticated mode allows one to reduce the time between when a
timestamp is obtained by a sender and when the packet is shipped out.
(Encryption causes extra delay, which leads to inaccuracy.)
- All of OWAMP-Test AES Session-key, OWAMP-Control AES
Session-key, and the SID are comprised of 16 octets. OWAMP-Test
HMAC Session-key and OWAMP-Control HMAC Session-key are comprised of
32 octets.
- After nine years of IPsec, only 0.05%
of traffic on an advanced backbone network, such as Abilene, uses
IPsec (for comparison purposes with encryption above layer 4, SSH
use is at 2-4% and HTTPS use is at 0.2-0.6%).
For numerous reasons in Section 6.6,
we decided to use a simple cryptographic
protocol (based on a block cipher in CBC mode) that is
different from
TLS and IPsec.
- A Two-way Active Measurement Protocol (TWAMP)
draft-ietf-ippm-twamp-00
- OWAMP @ Internet2
- OWAMP Performance and
Interoperability Tests [PDF]
- Plug-and-Play (Network) Performance
Monitoring
Reference Slides
- Better Illustration of E2E
piPEs in 2003.2.4 Miami Joint Tech BOF
- 2003.5.15. Almost the same presentation as
the previous one.
- Network Measurement Tools Summary of ESnet, 2006.02.
- Presentation of IEPM Group in
ESCC 2003.02 Miami Meeeting. It appeals that a common monitoring
infrastructure is necessary. See http://www-iepm.slac.stanford.edu
for more tools developed by IEPM.
- IETF IPPM WG Meeting in 2006.03
- Some Management and
Diagnostic Tools mentioned by UCLA NOC.
- 2005.10 Active and Passive Measurements
by OARnet
- Active Measurements
- Require injecting test packets into the network to determine network
topology or end-to-end performance of network paths
- +)Better characterize end-user perceived application-quality since
they emulate experience of actual end-application traffic using a
few test packets
- -)They consume bandwidth required by actual application traffic
- Passive Measurements
Do not inject test packets in the network but require capturing of
packets and their corresponding timestamps transmitted by
applications running on network-attached devices over various
network links
-
+)Do not inject test traffic and data is obtained from devices that
are involved in the functioning of the network
-
-)They impose large overhead on network devices to keep track of
such information in addition to their core functionality of
forwarding packets
- "Delay" is a metric for active measurement, not passive
measurement.
- Ohio State developed their own toolkit:
ActiveMon
- H.323
Beacon is recommended as "An easy to install and use
tool that is open source" to evaluate video/audio stream MOS values.
- 2006.4.5
OWAMP and BWCTL: Installation and Configuration by Jeff
Boote (boote@internet2.edu) @
Network Performance Workshop
- Hyperlinks to the cookbook of OWAMP and BWCTL are given.
- Detailed configuration steps are illustrated.
- 2006.5.29 Jeff, Internet2 E2EPI
- By defining clean APIs and protocols between each layer,
the diagnostic costs can be reduced.
- Users can access the statistics of network data from
Abilene
Observatory. TANet can try to establish similar mechanism to
allow researchers requesting similar data from TANet.
- Encourage people to deploy open source tools BWCTL/Iperf,
OWAMP, NDT.
- Set DSCP (a.k.a. first 6 bits of the TOS byte).
- The Performance Bottleneckpplication, Computer, or
Network
- Measurement Tools are VERY important!
- Case Study at SC'04
- Team spent over 1 hour looking at Win XP config, trying to verify
Buffer size
- Single NDT test verified this in under 30 seconds
- Full to Half mismatch detection - Asymmetric throughput
(opposite direction is less)
- Three major tools:
- BWCTL (Bandwidth Control wrapper for NLANR Iperf)
- OWAMP: One-Way Active Measurement
- NDT: Network Diagnostic Tool
- User-Perceived
Performance Measurement on the Internet
- This talk addresses several network measurement projects:
- NIMI (National Internet Measurement Infrastructure)
- E2E piPEs (of course)
- Proxy-Based Tools: Liston Proxy / Medusa Proxy
- AMP
- Network Weather Service
- Because the Internet was designed following the end-to-end
principle, end-to-end performance is ultimately the most meaningful
to measure
- Network users see end-to-end performance directly and could be
effectively used as monitor points
Podcast
-
Five Critical Issues for VoIP Implementation.
- What to do with your existing device
- Assessing your network status
- Redesign your network
- Pilot Strategy
- Training
Papers for Further Reading
- P Calyam, D Krymskiy, M Sridharan, P Schopis,
"Active and passive measurements on campus, regional and national
network backbone paths",
Computer Communications and Networks, 2005. ICCCN 2005
- R Kwitt, T Fichtel, T Pfeiffenberger,
"Measuring Perceptual VoIP Speech Quality over UMTS
- F Michaut, F Lepage,
"Application-oriented network metrology: metrics and active
measurement tools",
Communications Surveys & Tutorials, IEEE,
2005
- A Network Monitoring System with a
Peer-to-Peer Architecture
- Broom, S.R.: ‘VoIP quality assessment: taking account of the
edgedevice’,
IEEE Trans. Audio, Speech and Language Processing, Volume
14, Issue 6, Nov. 2006, pp. 1977–1983.
- “RTP Control protocol Extended Reports (RTCP XR)”, IETF RFC 3611,
Nov. 2003.
- L. Carvalho, J. Angeja and A. Navarro: ‘A new packet loss model of
the
IEEE 802.11g wireless network for multimedia communications’, IEEE
Trans. on Consumer Electronics, Volume 51, Issue 3, Aug. 2005, pp.
809-814.
- Conformance testing for voice over IP transmission quality
assessment
models, ITU-T, P.564, Nov. 2007.
- More ...