Transcript Cross-layer Monitoring in IPTV Network

“Group 0”
Lei Zhang 3421970
Zhenyu Li 3429940
Peng Zhou 3429310
Xiaohang Yin 3434340
Mauricio Barreto 3358630
Nilram Azadpeyma 3371520
TELE 9752 – Network
Operations & Control
Group Presentation
Week 12 – S2, 2013
1






Introduction
Aims and Challenges - Why
Monitoring Metrics - What
In-network & Client-side Monitoring – How &
Where
The ALICANTE Monitoring Approach
Conclusion and Our Opinion
2
IPTV: Delivery of multimedia content over IP-based networks
with guaranteed quality of service – paid service.
•Why IPTV business is so attractive?
40
35
30
25
revenue
billion
20
15
10
5
0
2009
2013
• IPTV must compete with traditional TV delivery systems.
• Very high reliability and QoS/QoE.
3
content
content
service
service
network
network
IPTV
Traditional TV
To achieve comparable QoS/QoE IPTV needs:
Integrated service &
network management
Real-time monitoring
• Monitoring must cover all layers in all system segments.
• Cross-layer monitoring is essential!
4

Hierarchical structure
• A Super Head-end
• Local Video Head-ends
• Core/Edge Networks
• Access Multiplexer
• Customer Gateway
• Presentation Devices
Service Assurance for IPTV
Internet Access Platform
Cellular Phone Network
Television Service
Figure 1. Typical architecture of an IPTV network. [7]
• Outage could be common
• Short outage is not tolerable
Even Small
Impairments
should be detected
and mitigated.
5

Reactive
 Response of the monitoring system to an
abnormal situation.

Proactive
 Behaviour of the monitoring system under
normal operation.
6

Reactive (Fault Management)
Detection of service outage
• Even packet losses on the order of 0.1% can cause
perceptible distortions to video and audio.
Estimation of problem
• Small losses or increased jitter can degrade the perceived
quality.
Localization of the failure point
• Impact of the problem depends on the point of failure.
• Failure in access multiplexer Vs. super head-end.
Assessment of the impact
• Map this event to its actual impact on viewer perceived
quality.
7

Proactive
Prevention of failure/outage (P)
• Achieved by checking the resource utilization and the
workload of the system components and identifying system
bottlenecks.
Detection of SLA status
• End-to-end monitoring data can be used.
Monitoring user behavior (P)
• To perform long-term resource planning.
• Optimize system operations.
Integrate monitoring with other components
• Complete picture of the status.
8
Scalability
Monitoring Core,
Edge & Access
Networks
Interoperability
• Vast number of involved
devices  Huge volume of
reports  Filter and process
• Service delivered may belong
to different network operators.
Willing to expose network
monitoring data? (Not really!)
• Different management
systems
9
Cross-layer Monitoring
 Parameters monitored in IPTV network:

 User/QoE Metrics
 User Perception
 Application Metrics
 During Decoding and Presentation
 Transport/Network Metrics
 Performance of the Corresponding Protocols
 Other:
 Protocol-Specific Metrics
FEC
 User Behavior Metrics
10



Within the Distribution Network, Core, Edge, and Access.
Needs to be distinguished from overall network monitoring.
Mainly collects:



Transport/Network Metrics e.g. Packet Loss and Jitter
Packet filtering and sampling is often used due to huge
traces.
Retrieval of measured metrics:

SNMP, IPFIX（Avoid overhead of SNMP/syslog)
 Modern Routers and Switches

Higher-Layer Metrics

Application/Stream
 Probe Modules
○
○
Number of Probes: Detection of Issue Vs. Cost and Complexity
Active Monitoring  Imposed Load /Traffic on the Network
11

Performed at 3 points:
Customer Network Gateway
Decoder/STB
Presentation Device
12
Figure. Architecture of the ALICANTE monitoring subsystem.［7］
13
Maximization of QoS and QoE of media
services is achieved.

a) via dynamic adaptation of media streams, taking
place in the network and also at the edges.

b) via content-aware mechanisms in the network,
performing automatic service recognition and
differentiation.
14

Four sets of cross-layer metrics are
identified corresponding to different
elements:
Metrics

Source
Host Metrics
Content Server & Home Box
VCAN Metrics
Network
Session Metrics
Home Box & User Terminals
Application/ QoE Metrics
User Terminals
Key feature: Decentralized Monitoring.
15

Thorough monitoring can only be achieved through cross-layer
monitoring.

All domains involved in IPTV service provisioning must be
monitored.

The Alicante monitoring approach is a practical implementation
of the principles of end to end service monitoring.

Decentralized monitoring allows adaptive management and
promotes collaboration among actors and across domains.
16

Lack of experimental support for ideas presented in
paper. The paper is more an opinion piece than a
research paper.

Conclusions should include: Use of virtual networks
allows data to be shared without exposing
commercially sensitive information.
17
[1] “IPTV Global Forecast — 2009 to 2013, Semiannual IPTV Global
Forecast Report,” Nov. 2009, MRG Group Inc.
[2] ITU-T Rec. G.1081, “Performance Monitoring Points for IPTV,” Oct.
2008.
[3] L. Boula, H. Koumaras, and A. Kourtis, “An Enhanced IMS
Architecture Featuring Cross-Layer Monitoring and Adaptation
Mechanisms,” ICAS ’09), Valencia, Spain, Apr. 20–25, 2009.
[4] ITU-T Rec. G.1080, “Quality of Experience Requirements for IPTV
Services,” Dec. 2008.
[5] ITU-T Rec. J.241, “Quality of Service Ranking and Mea- surement
Methods for Digital Video Services Delivered over Broadband IP
Networks,” Apr. 2005.
[6] ETSI TR 101 290 V1.2.1, “Digital Video Broadcasting (DVB);
Measurement Guidelines for DVB Systems,” tech. rep., May 2001.
[7] Boula L, Gardikis G, Xilouris G, et al. Cross-layer Monitoring in IPTV
Networks[J].
18
19