Multimedia Quality of Service What is it Multimedia

Multimedia, Quality of Service: What is it? Multimedia applications: network audio and video (“continuous media”) Qo. S network provides application with level of performance needed for application to function. Qo. S Networking 1

Goals Principles q Classify multimedia applications q Identify the network services the apps need q Making the best of best effort service q Mechanisms for providing Qo. S Protocols and Architectures q Specific protocols for best-effort q Architectures for Qo. S Networking 2

Chapter 7 outline q 7. 1 Multimedia Networking Applications q 7. 2 Streaming stored audio and video q 7. 3 Real-time Multimedia: Internet Phone study q 7. 6 Beyond Best Effort q 7. 7 Scheduling and Policing Mechanisms q 7. 8 Integrated Services and Differentiated Services Qo. S Networking 3

MM Networking Applications Classes of MM applications: 1) Streaming stored audio and video 2) Streaming live audio and video 3) Real-time interactive audio and video Fundamental characteristics: q Typically delay sensitive o o end-to-end delay jitter q But loss tolerant: infrequent losses cause minor glitches q Antithesis of data, which are loss intolerant but delay tolerant. Jitter is the variability of packet delays within the same packet stream Qo. S Networking 4

Streaming Stored Multimedia Streaming: q media stored at source q transmitted to client q streaming: client playout begins before all data has arrived q timing constraint for still-to-be transmitted data: in time for playout Qo. S Networking 5

Cumulative data Streaming Stored Multimedia: What is it? 1. video recorded 2. video sent network delay 3. video received, played out at client time streaming: at this time, client playing out early part of video, while server still sending later part of video Qo. S Networking 6

Streaming Stored Multimedia: Interactivity q VCR-like functionality: client can pause, rewind, FF, push slider bar o 10 sec initial delay OK o 1 -2 sec until command effect OK o RTSP often used (more later) q timing constraint for still-to-be transmitted data: in time for playout Qo. S Networking 7

Streaming Live Multimedia Examples: q Internet radio talk show q Live sporting event Streaming q playback buffer q playback can lag tens of seconds after transmission q still have timing constraint Interactivity q fast forward impossible q rewind, pause possible! Qo. S Networking 8

Interactive, Real-Time Multimedia q applications: IP telephony, video conference, distributed interactive worlds q end-end delay requirements: o audio: < 150 msec good, < 400 msec OK • includes application-level (packetization) and network delays • higher delays noticeable, impair interactivity q session initialization o how does callee advertise its IP address, port number, encoding algorithms? Qo. S Networking 9

Multimedia Over Today’s Internet TCP/UDP/IP: “best-effort service” q no guarantees on delay, loss ? ? ? But you said multimedia apps requires ? Qo. S and level of performance to be ? ? effective! ? ? Today’s Internet multimedia applications use application-level techniques to mitigate (as best possible) effects of delay, loss Qo. S Networking 10

How should the Internet evolve to better support multimedia? Integrated services philosophy: q Fundamental changes in Internet so that apps can reserve end-to-end bandwidth q Requires new, complex software in hosts & routers Laissez-faire q no major changes q more bandwidth when needed q content distribution, application -layer multicast o Differentiated services philosophy: q Fewer changes to Internet infrastructure, yet provide 1 st and 2 nd class service. application layer Qo. S Networking 11

A few words about audio compression q Analog signal sampled at constant rate o o telephone: 8, 000 samples/sec CD music: 44, 100 samples/sec q Each sample quantized, i. e. , rounded o e. g. , 28=256 possible quantized values q Each quantized value represented by bits o q Example: 8, 000 samples/sec, 256 quantized values --> 64, 000 bps q Receiver converts it back to analog signal: o some quality reduction Example rates q CD: 1. 411 Mbps q MP 3: 96, 128, 160 kbps q Internet telephony: 5. 3 13 kbps 8 bits for 256 values Qo. S Networking 12

A few words about video compression q Video is sequence of images displayed at constant rate o e. g. 24 images/sec q Digital image is array of pixels q Each pixel represented by bits q Redundancy o o spatial temporal Examples: q MPEG 1 (CD-ROM) 1. 5 Mbps q MPEG 2 (DVD) 3 -6 Mbps q MPEG 4 (often used in Internet, < 1 Mbps) Research: q Layered (scalable) video o adapt layers to available bandwidth Qo. S Networking 13

Chapter 7 outline q 7. 1 Multimedia Networking Applications q 7. 2 Streaming stored audio and video q 7. 3 Real-time Multimedia: Internet Phone study q 7. 6 Beyond Best Effort q 7. 7 Scheduling and Policing Mechanisms q 7. 8 Integrated Services and Differentiated Services Qo. S Networking 14

Streaming Stored Multimedia Application-level streaming techniques for making the best out of best effort service: o client side buffering o use of UDP versus TCP o multiple encodings of multimedia Media Player q jitter removal q decompression q error concealment q graphical user interface w/ controls for interactivity Qo. S Networking 15

Internet multimedia: simplest approach q audio or video stored in file q files transferred as HTTP object o o received in entirety at client then passed to player audio, video not streamed: q no, “pipelining, ” long delays until playout! Qo. S Networking 16

Internet multimedia: streaming approach q browser GETs metafile q browser launches player, passing metafile q player contacts server q server streams audio/video to player Qo. S Networking 17

Streaming from a streaming server q This architecture allows for non-HTTP protocol between server and media player q Can also use UDP instead of TCP. Qo. S Networking 18

constant bit rate video transmission variable network delay client video reception constant bit rate video playout at client buffered video Cumulative data Streaming Multimedia: Client Buffering time client playout delay q Client-side buffering, playout delay compensate for network-added delay, delay jitter Qo. S Networking 19

Streaming Multimedia: Client Buffering constant drain rate, d variable fill rate, x(t) buffered video q Client-side buffering, playout delay compensate for network-added delay, delay jitter Qo. S Networking 20

Streaming Multimedia: UDP or TCP? UDP q server sends at rate appropriate for client (oblivious to network congestion !) o often send rate = encoding rate = constant rate o then, fill rate = constant rate - packet loss q short playout delay (2 -5 seconds) to compensate for network delay jitter q error recover: time permitting TCP q send at maximum possible rate under TCP q fill rate fluctuates due to TCP congestion control q larger playout delay: smooth TCP delivery rate q HTTP/TCP passes more easily through firewalls Qo. S Networking 21

Streaming Multimedia: client rate(s) 1. 5 Mbps encoding 28. 8 Kbps encoding Q: how to handle different client receive rate capabilities? o 28. 8 Kbps dialup o 100 Mbps Ethernet A: server stores, transmits multiple copies of video, encoded at different rates Qo. S Networking 22

User Control of Streaming Media: RTSP HTTP q Does not target multimedia content q No commands for fast forward, etc. RTSP: RFC 2326 q Client-server application layer protocol. q For user to control display: rewind, fast forward, pause, resume, repositioning, etc… What it doesn’t do: q does not define how audio/video is encapsulated for streaming over network q does not restrict how streamed media is transported; it can be transported over UDP or TCP q does not specify how the media player buffers audio/video Qo. S Networking 23

RTSP: out of band control FTP uses an “out-of-band” control channel: q A file is transferred over one TCP connection. q Control information (directory changes, file deletion, file renaming, etc. ) is sent over a separate TCP connection. q The “out-of-band” and “inband” channels use different port numbers. RTSP messages are also sent out-of-band: q RTSP control messages use different port numbers than the media stream: out -of-band. o Port 554 q The media stream is considered “in-band”. Qo. S Networking 24

RTSP Example Scenario: q metafile communicated to web browser q browser launches player q player sets up an RTSP control connection, data connection to streaming server Qo. S Networking 25

Metafile Example <title>Twister</title> <session> <group language=en lipsync> <switch> <track type=audio e="PCMU/8000/1" src = "rtsp: //audio. example. com/twister/audio. en/lofi"> <track type=audio e="DVI 4/16000/2" pt="90 DVI 4/8000/1" src="rtsp: //audio. example. com/twister/audio. en/hifi"> </switch> <track type="video/jpeg" src="rtsp: //video. example. com/twister/video"> </group> </session> Qo. S Networking 26

RTSP Operation Qo. S Networking 27

RTSP Exchange Example C: SETUP rtsp: //audio. example. com/twister/audio RTSP/1. 0 Transport: rtp/udp; compression; port=3056; mode=PLAY S: RTSP/1. 0 200 1 OK Session 4231 C: PLAY rtsp: //audio. example. com/twister/audio. en/lofi RTSP/1. 0 Session: 4231 Range: npt=0 C: PAUSE rtsp: //audio. example. com/twister/audio. en/lofi RTSP/1. 0 Session: 4231 Range: npt=37 C: TEARDOWN rtsp: //audio. example. com/twister/audio. en/lofi RTSP/1. 0 Session: 4231 S: 200 3 OK Qo. S Networking 28

Chapter 7 outline q 7. 1 Multimedia Networking Applications q 7. 2 Streaming stored audio and video q 7. 3 Real-time Multimedia: Internet Phone case study q 7. 6 Beyond Best Effort q 7. 7 Scheduling and Policing Mechanisms Qo. S Networking 29

Real-time interactive applications q PC-2 -PC phone o instant messaging services are providing this q PC-2 -phone Dialpad o Net 2 phone q videoconference with Webcams o Qo. S Networking 30

Interactive Multimedia: Internet Phone q speaker’s audio: alternating talk spurts, silent periods. o 64 kbps during talk spurt q pkts generated only during talk spurts o 20 msec chunks at 8 Kbytes/sec: 160 bytes data q application-layer header added to each chunk. q Chunk+header encapsulated into UDP segment. q application sends UDP segment into socket every 20 msec during talkspurt. Qo. S Networking 31

Internet Phone: Packet Loss and Delay q network loss: IP datagram lost due to network congestion (router buffer overflow) q delay loss: IP datagram arrives too late for playout at receiver o o delays: processing, queueing in network; end-system (sender, receiver) delays typical maximum tolerable delay: 400 ms q loss tolerance: depending on voice encoding, losses concealed, packet loss rates between 1% and 10% can be tolerated. Qo. S Networking 32

constant bit rate transmission variable network delay (jitter) client reception constant bit rate playout at client buffered data Cumulative data Delay Jitter client playout delay time q Consider the end-to-end delays of two consecutive packets: difference can be more or less than 20 msec Qo. S Networking 33

Internet Phone: Fixed Playout Delay q Receiver attempts to playout each chunk exactly q msecs after chunk was generated. o chunk has time stamp t: play out chunk at t+q. o chunk arrives after t+q: data arrives too late for playout, data “lost” q Tradeoff for q: o large q: less packet loss o small q: better interactive experience Qo. S Networking 34

Fixed Playout Delay • Sender generates packets every 20 msec during talk spurt. • First packet received at time r • First playout schedule: begins at p • Second playout schedule: begins at p’ Qo. S Networking 35

Adaptive Playout Delay, I q Goal: minimize playout delay, keeping late loss rate low q Approach: adaptive playout delay adjustment: o Estimate network delay, adjust playout delay at beginning of each talk spurt. o Silent periods compressed and elongated. o Chunks still played out every 20 msec during talk spurt. Dynamic estimate of average delay at receiver: where u is a fixed constant (e. g. , u =. 01). Qo. S Networking 36

Adaptive playout delay II Also useful to estimate the average deviation of the delay, vi : The estimates di and vi are calculated for every received packet, although they are only used at the beginning of a talk spurt. For first packet in talk spurt, playout time is: where K is a positive constant. Remaining packets in talkspurt are played out periodically Qo. S Networking 37

Adaptive Playout, III Q: How does receiver determine whether packet is first in a talkspurt? q If no loss, receiver looks at successive timestamps. o difference of successive stamps > 20 msec -->talk spurt begins. q With loss possible, receiver must look at both time stamps and sequence numbers. o difference of successive stamps > 20 msec and sequence numbers without gaps --> talk spurt begins. Qo. S Networking 38

Recovery from packet loss (1) forward error correction (FEC): q Playout delay needs to be simple scheme fixed to the time to receive q for every group of n chunks all n+1 packets create a redundant chunk by q Tradeoff: exclusive OR-ing the n original o increase n, less chunks bandwidth waste q send out n+1 chunks, increasing o increase n, longer playout the bandwidth by factor 1/n. delay q can reconstruct the original n o increase n, higher chunks if there is at most one probability that 2 or lost chunk from the n+1 chunks more chunks will be lost Qo. S Networking 39

Recovery from packet loss (2) 2 nd FEC scheme • “piggyback lower quality stream” • send lower resolution audio stream as the redundant information • for example, nominal stream PCM at 64 kbps and redundant stream GSM at 13 kbps. • Whenever there is non-consecutive loss, the receiver can conceal the loss. • Can also append (n-1)st and (n-2)nd low-bit rate chunk Qo. S Networking 40

Recovery from packet loss (3) Interleaving q chunks are broken up into smaller units q for example, 4 5 msec units per chunk q Packet contains small units from different chunks q if packet is lost, still have most of every chunk q has no redundancy overhead q but adds to playout delay Qo. S Networking 41

Summary: Internet Multimedia: bag of tricks q use UDP to avoid TCP congestion control (delays) for time- sensitive traffic q client-side adaptive playout delay: to compensate for delay q server side matches stream bandwidth to available client-to- server path bandwidth o o chose among pre-encoded stream rates dynamic server encoding rate q error recovery (on top of UDP) o o o FEC, interleaving retransmissions, time permitting conceal errors: repeat nearby data Qo. S Networking 42

Chapter 7 outline q 7. 1 Multimedia Networking Applications q 7. 2 Streaming stored audio and video q 7. 3 Real-time Multimedia: Internet Phone study q 7. 6 Beyond Best Effort q 7. 7 Scheduling and Policing Mechanisms q 7. 8 Integrated Services and Differentiated Services Qo. S Networking 43

Improving QOS in IP Networks Thus far: “making the best of best effort” Future: next generation Internet with Qo. S guarantees o RSVP: signaling for resource reservations o Differentiated Services: differential guarantees o Integrated Services: firm guarantees q simple model for sharing and congestion studies: Qo. S Networking 44

Principles for QOS Guarantees q Example: 1 Mbps IP phone, FTP share 1. 5 Mbps link. o bursts of FTP can congest router, cause audio loss o want to give priority to audio over FTP Principle 1 packet marking needed for router to distinguish between different classes; and new router policy to treat packets accordingly Qo. S Networking 45

Principles for QOS Guarantees (more) q what if applications misbehave (audio sends higher than declared rate) o policing: force source adherence to bandwidth allocations q marking and policing at network edge: o similar to ATM UNI (User Network Interface) Principle 2 provide protection (isolation) for one class from others Qo. S Networking 46

Principles for QOS Guarantees (more) q Allocating fixed (non-sharable) bandwidth to flow: inefficient use of bandwidth if flow doesn’t use its allocation Principle 3 While providing isolation, it is desirable to use resources as efficiently as possible Qo. S Networking 47

Principles for QOS Guarantees (more) q Basic fact of life: can not support traffic demands beyond link capacity Principle 4 Call Admission: flow declares its needs, network may block call (e. g. , busy signal) if it cannot meet needs Qo. S Networking 48

Summary of Qo. S Principles Let’s next look at mechanisms for achieving this …. Qo. S Networking 49

Chapter 7 outline q 7. 1 Multimedia Networking Applications q 7. 2 Streaming stored audio and video q 7. 3 Real-time Multimedia: Internet Phone study q 7. 6 Beyond Best Effort q 7. 7 Scheduling and Policing Mechanisms q 7. 8 Integrated Services and Differentiated Services Qo. S Networking 50

Scheduling And Policing Mechanisms q scheduling: choose next packet to send on link q FIFO (first in first out) scheduling: send in order of arrival to queue o o real-world example? discard policy: if packet arrives to full queue: who to discard? • Tail drop: drop arriving packet • priority: drop/remove on priority basis • random: drop/remove randomly Qo. S Networking 51

Scheduling Policies: more Priority scheduling: transmit highest priority queued packet q multiple classes, with different priorities o o class may depend on marking or other header info, e. g. IP source/dest, port numbers, etc. . Real world example? Qo. S Networking 52

Scheduling Policies: still more round robin scheduling: q multiple classes q cyclically scan class queues, serving one from each class (if available) q real world example? Qo. S Networking 53

Scheduling Policies: still more Weighted Fair Queuing: q generalized Round Robin q each class gets weighted amount of service in each cycle Qo. S Networking 54

Policing Mechanisms Goal: limit traffic to not exceed declared parameters Three common-used criteria: q (Long term) Average Rate: how many pkts can be sent per unit time (in the long run) o crucial question: what is the interval length: 100 packets per sec or 6000 packets per min have same average! q Peak Rate: e. g. , 6000 pkts per min. (ppm) avg. ; 1500 ppm peak rate q (Max. ) Burst Size: max. number of pkts sent consecutively (with no intervening idle) Qo. S Networking 55

Policing Mechanisms Token Bucket: limit input to specified Burst Size and Average Rate. q bucket can hold b tokens q tokens generated at rate r token/sec unless bucket full q over interval of length t: number of packets admitted less than or equal to (r t + b). Qo. S Networking 56

Policing Mechanisms (more) q token bucket, WFQ combine to provide guaranteed upper bound on delay, i. e. , Qo. S guarantee! arriving traffic token rate, r bucket size, b WFQ per-flow rate, R D = b/R max Qo. S Networking 57

Chapter 7 outline q 7. 1 Multimedia Networking Applications q 7. 2 Streaming stored audio and video q 7. 3 Real-time Multimedia: Internet Phone study q 7. 6 Beyond Best Effort q 7. 7 Scheduling and Policing Mechanisms q 7. 8 Integrated Services and Differentiated Services Qo. S Networking 58

IETF Integrated Services q architecture for providing QOS guarantees in IP networks for individual application sessions q resource reservation: routers maintain state info of allocated resources, Qo. S req’s q admit/deny new call setup requests Question: can newly arriving flow be admitted with performance guarantees while not violated Qo. S guarantees made to already admitted flows? Qo. S Networking 59

Intserv: Qo. S guarantee scenario q Resource reservation o call setup, signaling (RSVP) o traffic, Qo. S declaration o per-element admission control request/ reply o Qo. S-sensitive scheduling (e. g. , WFQ) Qo. S Networking 60

Call Admission Arriving session must : q declare its QOS requirement R-spec: defines the QOS being requested q characterize traffic it will send into network o T-spec: defines traffic characteristics q signaling protocol: needed to carry R-spec and T-spec to routers (where reservation is required) o RSVP o Qo. S Networking 61

Intserv Qo. S: Service models [rfc 2211, rfc 2212] Guaranteed service: q worst case traffic arrival: leaky- Controlled load service: q "a quality of service closely bucket-policed source approximating the Qo. S that q simple (mathematically provable) same flow would receive bound on delay [Parekh 1992, Cruz 1988] from an unloaded network element. " arriving traffic token rate, r bucket size, b WFQ per-flow rate, R D = b/R max Qo. S Networking 62

IETF Differentiated Services Concerns with Intserv: q Scalability: signaling, maintaining per-flow router state difficult with large number of flows q Flexible Service Models: Intserv has only two classes. Also want “qualitative” service classes o o “behaves like a wire” relative service distinction: Platinum, Gold, Silver Diffserv approach: q simple functions in network core, relatively complex functions at edge routers (or hosts) q Don’t define service classes, provide functional components to build service classes Qo. S Networking 63

Diffserv Architecture Edge router: r marking scheduling q per-flow traffic management q marks packets as in-profile and out-profile b . . . Core router: q per class traffic management q buffering and scheduling based on marking at edge q preference given to in-profile packets q Implements specified per-hop behavior (PHB) Qo. S Networking 64

Edge-router Packet Marking q profile: pre-negotiated rate A, bucket size B q packet marking at edge based on per-flow profile Rate A B User packets Possible usage of marking: q class-based marking: packets of different classes marked differently q intra-class marking: conforming portion of flow marked differently than non-conforming one Qo. S Networking 65

Classification and Conditioning q Packet is marked in the Type of Service (TOS) in IPv 4, and Traffic Class in IPv 6 q 6 bits used for Differentiated Service Code Point (DSCP) and determine PHB that the packet will receive q 2 bits are currently unused Qo. S Networking 66

Classification and Conditioning may be desirable to limit traffic injection rate of some class: q user declares traffic profile (e. g. , rate, burst size) q traffic metered, shaped if non-conforming Qo. S Networking 67

Forwarding (PHB) q PHB result in a different observable (measurable) forwarding performance behavior q PHB does not specify what mechanisms to use to ensure required PHB performance behavior q Examples: o o Class A gets x% of outgoing link bandwidth over time intervals of a specified length Class A packets leave first before packets from class B Qo. S Networking 68

Forwarding (PHB) PHBs being developed: q Expedited Forwarding: pkt departure rate of a class equals or exceeds specified rate o logical link with a minimum guaranteed rate q Assured Forwarding: 4 classes of traffic o each guaranteed minimum amount of bandwidth o each with three drop preference partitions Qo. S Networking 69

Multimedia Networking: Summary q multimedia applications and requirements q making the best of today’s best effort service q scheduling and policing mechanisms q next generation Internet: Intserv, RSVP, Diffserv Qo. S Networking 70