Transcript QoS and Video Delivery

QoS and Video Delivery
Presented by Wei Wei
Internet QoS: A Big Picture[1][2][3][4]
• Current Internet approach
Best effort
No guarantees
• Need of providing QoS for video
applications
Introduction
• Integrated Services (InteServ) and Resource
Reservation Protocol (RSVP)
• Differentiated Service (DiffServ)
• Multiprotocol Label Switching (MPLS)
• Traffic Engineering and Constraint Based
Routing
InteServ and RSVP
• Philosophy Behind
Routers have to be able to reserve resources to provide
special QoS for specific user packet streams
• Four components of IntServ Model
The signaling protocol (e.g. RSVP)
The admission control routine
The classifier
The packet scheduler
InteServ and RSVP
• How RSVP works?
InteServ and RSVP
• Sender sends a PATH Message to the receiver
specifying the characteristics of the traffic
Every intermediate router along the path forwards
the PATH Message to the next hop determined by
the routing protocol
• The receiver responds with a RESV Message to
request resources for the flow
InteServ and RSVP
• Every intermediate router along the path can reject
or accept the request of the RESV Message
Rejected, the router send an error message to the
receiver, the signaling process will terminate
 Accepted, Resources (link bandwidth and buffer
space) are allocated for the flow and the related flow
state information will be installed in the router
InteServ and RSVP
• Admission control
Decide whether a request for resources can be granted
• Classifier
When a router receives a packet, the classifier will perform
a Multi-Field (MF) classification and put the packet in a
specific queue based on the classification result
• Packet scheduler
Schedule the packet accordingly to meet its QoS
requirements
InteServ and RSVP
• Problems
– Not scalable
• Huge storage and processing overhead on the
routers
• The amount of state information increases
proportionally with the number of flows
– Requirement on routers is high
• All routers must implement RSVP, admission control, MF
classification, and packet scheduling
DiffServ
• Main Idea
– Push all complex functions to the edge routers
• Edge routers are responsible for multi-field
classification and translating that into a DSCP
• Core routers treat packets through DS field
– Marking the DS fields of packets differently,
then handling packets based on their DS fields
DiffServ
• Architecture
DiffServ
• Architecture
– Ingress routers
• Police/shape traffic
• Set Differentiated Service Code Point (DSCP) in
Diffserv (DS) field
– Core routers
• Implement Per Hop Behavior (PHB) for each DSCP
• Process packets based on DSCP
DiffServ
• Two types of service
– Assured service
– Premium service
• Plus, best-effort service
DiffServ
• Assured Service
– Defined in terms of user profile, how much assured
traffic is a user allowed to inject into the network
– Network: provides a lower loss rate than best-effort
• In case of congestion best-effort packets are dropped first
– User: sends no more assured traffic than its profile
• If it sends more, the excess traffic is converted to best-effort
DiffServ
• Premium service
– Provides the abstraction of a virtual pipe
between an ingress and an egress router
– Network: guarantees that premium packets are
not dropped and they experience low delay
– User: does not send more than the size of the
pipe
• If it sends more, excess traffic is delayed, and
dropped when buffer overflows
DiffServ
• Advantage
– Scalable
• Edge routers maintain per aggregate state
• Core routers maintain state only for a few traffic classes
– Easy implementation
• Incremental deployment is possible for Assured Forwarding
• Disadvantage
– Provide weaker service than InteServ
MPLS
• Label Switching
– Header of the packet contains a label that is
used to advance the packet toward its
destination
– The label simplifies the forwarding decision a
node must make for the packet
• A group of packets forwarded in the same
manner are said to belong to the same
Forwarding Equivalence Class (FEC)
MPLS
• Label Switched Paths (LSPs)
– Within an MPLS domain, a path is set up for a
given packet to travel based on a Forwarding
Equivalence Class (FEC)
– The LSP is set up prior to data transmission
MPLS
• MPLS improves packet forwarding performance
– Enhances and simplifies packet forwarding through
routers that use layer-2 switching
– Simplicity allows for easy implementation
• MPLS supports QoS for service differentiation
– Use traffic-engineered path set-up and support QoS
guarantees
– Classification and QoS service are determined by the
labels
Traffic Engineering and
Constraint-Based Routing
• Traffic Engineering is the process of
arranging how traffic flows through the
network so that congestion caused by
uneven network utilization can be avoided
• Constraint-Based Routing is used to
compute routes that are subject to multiple
constraints.
Traffic Engineering
• Network congestion can be caused by lack of
network resources or by uneven distribution of
traffic
– In the first case, all routers and links are overloaded and
the only solution is to provide more resources by
upgrading the infrastructure
– In the second case, uneven traffic distribution can be
caused by the current Dynamic Routing protocols such
as RIP and OSPF, because they always select the
shortest paths to forward packets
– Traffic Engineering can be utilized to avoid congestion
or to provide graceful degradation in case of congestion
Constraint-Based Routing
• Constraint-Based Routing with DiffServ
– Select routes for flows so that their QoS
requirements are most likely to be met
• Constraint-Based Routing with RSVP
– Select the path for RSVP messages
• Constraint-Based Routing with MPLS
– MPLS is a forwarding scheme and constraintbased routing is a routing scheme
User-Oriented QoS in Packet
Video Delivery[5]
• focuses on how the quality of service of video
delivery is perceived by the end user
• What may affect the QoS
– Encoding: Aritifacts and Delays
• All lossy compression schemes both distort and delay the
signal
• Degradations come from quantization. Generally, the higher
the quantization step, the higher the degradation.
• The amount of delay introduced is related to the size of
encoding buffer. The bigger the buffer, the smoother the bit
rate may become, but it is at the expense of higher delay.
Trade-off
User-Oriented QoS in Packet
Video Delivery
– Transmission: Loss and Delay
• The quality degradation caused by data loss depends on the
importance of the lost information type.
• Spatial propagation: variable-length coding
• Temporal propagation: I- P- frame
• How to Improve QoS
– Encoder side:
• Adaptive Quantization: allocate more bits to more important
part.
• Syntatic protection: increase resynchronization point to reduce
spatial and temporal error propagation.
User-Oriented QoS in Packet
Video Delivery
– Layered coding: Video can be encoded into
different layers, and each layer has different
importance to the video quality.
• Unequal error protection: add more FEC to the more
important layer
• DiffServ Network: classify video signal into
different service class, and let the more important
data receive higher service
• Rate Adaptation: adapt to the varying channel
User-Oriented QoS in Packet
Video Delivery
• Network Adaptation
– ARQ: causes delay
– FEC: constant overhead
– Hybrid ARQ and FEC
• Decoder side
– Error Concealment Techniques: spatial
interpolation, temporal interpolation
User-Oriented QoS in Packet
Video Delivery
• How to measure QoS
– A popular metric is PSNR. Unfortunately, a
higher PSNR does not mean a higher quality.
– Metirc based on the properties of the human
visual system:
• metric based on a subjective rating function
obtained by psychovisual experiments
• metrics relying on a model of the human visual
system (Sarnoff JND Vision Model, moving pictures
quality metric (MPQM) and PDM )
User-Oriented QoS in Packet
Video Delivery
• The Perceptual Impact of MPEG-2 Rate and
Data Loss
– The Impact of Encoding Rate on Video Quality
• The perceptual video quality increases as the bit rate
increases, but it will saturate at some high bit rate
point
User-Oriented QoS in Packet
Video Delivery
User-Oriented QoS in Packet
Video Delivery
• The Impact of Data Loss on Video Quality
– The video quality first remains constant with
PLR. Then, beyond a certain PLR, the
perceptual quality drops fast.
User-Oriented QoS in Packet
Video Delivery
User-Oriented QoS in Packet
Video Delivery
• Joint Impact Analysis
– The relation between quality and the encoding
bit rate for a given nonzero PLR exhibits like
Fig. 17. Video quality first increases with the
average bit rate and then decreases after a
certain point.
User-Oriented QoS in Packet
Video Delivery
Dynamic Quality of Service Framework
for Video in Broadband Networks[6]
• Key points:
– Concept of Softness QoS
– Client QoS renegotiation
Dynamic Quality of Service Framework
for Video in Broadband Networks
• Softness refers to the ability of an application to
gracefully scale its performance using the
available network bandwidth.
• Appropriate QoS parameters specific to the
network layer are negotiated at connection setup.
• Support terminal-initiated renegotiation for
bandwidth on demand.
Dynamic Quality of Service Framework
for Video in Broadband Networks
• Advantages:
– Lower service blocking probability
– Achieve network efficiency, while maintaining
an acceptable application-level performance.
• Disadvantages:
– Increase setup delay
Dynamic Quality of Service Framework
for Video in Broadband Networks
• System Model
– Data Plane:
• Client:
– Buffer: stores packets until the decoder is ready to process them,
smooth the display rate.
– Decoding and display modules: error concealment module,
• Server
– Source rate control module: adjust video’s rate to match the
available bandwidth.
– Rate shaper: shapes the traffic to make sure that it fits the traffic
profile provided by the network
Dynamic Quality of Service Framework
for Video in Broadband Networks
– Control Plane:
• Client
– specify the terminal and the application requirements, and
set up “QoS contract”
• Server
– bandwidth renegotiation with the network control to
maintain the clients “QoS contract”
• Network
– Connection admission:
– Bandwidth (re)allocation:
QoS-Sensitive
[7]
Flows
• Concentrate on the issues and principles
concerning router modification to provide
QoS support in IP network
–
–
–
–
Packet marking
Packet classification
Scheduling and Queue Management
Traffic descriptions and Admission Control
A Management and Control Architecture for
Providing IP Differentiated Services in
MPLS-Based Networks[8]
• Current DiffServ approach concentrates on
control/data plane mechanisms to support
QoS, while this paper focuses on
management plane to support Diffservbased QoS.
A Management and Control Architecture for
Providing IP Differentiated Services in
MPLS-Based Networks
• Standalization of definition of Service-level
Specifications (SLS)
• Architecture
– SLS Management
• Subscribing and negotiating SLSs with users or
other peer ASs and performs admission control for
the dynamic request of subscribed SLSs
A Management and Control Architecture for
Providing IP Differentiated Services in
MPLS-Based Networks
– Traffic Engineering
• Selecting paths that are capable of meeting the QoS
requirements for a given traffic demand.
– Policy management
Conclusions
• How to Provide QoS in current Internet?
– Trying to make “some changes” to the current routers
to achieve some kind of QoS
• User-Oriented QoS
– There exists gap between the provision of networklevel QoS and the actual QoS requirements of
applications. This gap causes distributed multimedia
applications to inefficiently use network bandwidth,
leading to poor end-to-end system performance.
References
[1] X. Xiao, et al. “Internet QoS: A Big Picture”, IEEE
Network, Mar/April 1999
[2] L. Zhang, et al. "RSVP: A New Resource Reservation
Protocol", IEEE Communications Magazine, 31(9):8-18,
September 1993
[3] R. Braden, et al. "Integrated Services in the Internet
Architecture: an Overview", RFC 1633, June 1994
[4] K. Nichols , et al. "A Two-bit Differentiated Services
Architecture for the Internet",
References
[5] O. Verscheure, et al. “User-Oriented QoS in Packet Video Delivery”,
IEEE Network, Nov./Dec. 1998, pp. 12-21
[6] D. Reininger, et al. “A Dynamic Quality of Service Framework for
Video in Broadband Networks,” IEEE Network, Nov./Dec. 1998, pp.
22-34
[7] S.N. Bhatti, et al. “QoS-Sensitive Flows: Issues in IP Packet
Handling”, IEEE Internet Computing, Jul./Aug. 2000, pp. 48-57
[8] P. Trimintzios, et al. “A Management and Control Architecture for
Providing IP Differentiated Services in MPLS-Based Networks,” IEEE
Communications Magazine, May 2001, pp. 80-88