1 School of Computing Science Simon Fraser University CMPT 771/471: Internet Architecture and Protocols Multimedia Networking and Quality of Service Instructor: Dr. Mohamed Hefeeda

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

3 Chapter 7: Goals Principles r Classify multimedia applications r Identify the network services the apps need r Making the best of best effort service r Mechanisms for providing QoS Protocols and Architectures r Specific protocols for best-effort r Architectures for QoS

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

5 A few words about audio compression r Analog signal sampled at constant rate m telephone: 8,000 samples/sec m CD music: 44,100 samples/sec r Each sample quantized, i.e., rounded m e.g., 2 8 =256 possible quantized values r Each quantized value represented by bits m 8 bits for 256 values r Example: 8,000 samples/sec, 256 quantized values --> 64,000 bps r Receiver converts it back to analog signal: m some quality reduction Example rates r CD: Mbps r MP3: 96, 128, 160 kbps r Internet telephony: kbps

6 A few words about video compression r Video is sequence of images displayed at constant rate m e.g. 24 images/sec r Digital image is array of pixels r Each pixel represented by bits r Redundancy m spatial m temporal Examples: r MPEG 1 (CD-ROM) 1.5 Mbps r MPEG2 (DVD) 3-6 Mbps r MPEG4 (often used in Internet, < 1 Mbps) Research: r Layered (scalable) video m adapt layers to available bandwidth r Fine-grain scalable (FGS) video m Adapt at bit level

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

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

9 RTSP Example : r metafile communicated to web browser r browser launches player r player sets up an RTSP control connection, data connection to streaming server

10 Real-time interactive applications r PC-2-PC phone m instant messaging services are providing this r PC-2-phone m Dialpad m Net2phone r videoconference with Webcams Going to now look at a PC-2-PC Internet phone example in detail

11 Interactive Multimedia: Internet Phone r speaker’s audio: alternating talk spurts, silent periods. m 64 kbps during talk spurt r pkts generated only during talk spurts m 20 msec chunks at 8 Kbytes/sec: 160 bytes data r application-layer header added to each chunk r Chunk+header encapsulated into UDP segment r application sends UDP segment into socket every 20 msec during talk spurt r Need to handle packet losses and delay jitter

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

13 Recovery from packet loss (2) 2nd 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

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

15 Handling Delay Jitter r Receiver uses a playout buffer m Fixed buffer m Adaptive buffer: depends on network conditions

16 Adaptive Playout Buffer (1) Dynamic estimate of average delay at receiver: where u is a fixed constant (e.g., u =.01). r Goal: minimize playout delay, keeping late loss rate low r Approach: adaptive playout delay adjustment: m Estimate network delay, adjust playout delay at beginning of each talk spurt. m Silent periods compressed and elongated. m Chunks still played out every 20 msec during talk spurt.

17 Adaptive Playout Buffer (2) Also useful to estimate the average deviation of the delay, v i : The estimates d i and v i 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 talk spurt are played out periodically

18 Real-Time Protocol (RTP): FRC 1889 r RTP specifies packet structure audio and video data m payload type identification m packet sequence numbering m time stamping r RTP runs in the end systems r RTP packets are encapsulated in UDP segments r RTP does not provide any mechanism to ensure QoS m RTP encapsulation is only seen at the end systems

19 RTP Header Payload Type (7 bits): Indicates type of encoding currently being used: e.g., Payload type 0: PCM mu-law, 64 kbps Payload type 33, MPEG2 video Sequence Number (16 bits): Increments by one for each RTP packet sent, and may be used to detect packet Timestamp field (32 bytes long). Reflects the sampling instant of the first byte in the RTP data packet. SSRC field (32 bits long). Identifies the source of the RTP stream. Each stream in a RTP session should have a distinct SSRC.

20 Real-Time Control Protocol (RTCP) r Works in conjunction with RTP r Each participant in RTP session periodically transmits RTCP control packets to all other participants r Each RTCP packet contains sender and/or receiver reports m statistics, e.g., number of packets sent, number of packets lost, interarrival jitter, etc. m used to control app performance m Sender may modify its transmissions based on feedback

21 RTCP Packets Receiver report packets: r fraction of packets lost, last sequence number, average interarrival jitter. Sender report packets: r SSRC of the RTP stream, the current time, the number of packets sent, and the number of bytes sent. Source description packets: r address of sender, sender's name, SSRC of associated RTP stream. r Provide mapping between the SSRC and the user/host name.

22 RTCP: Synchronization of Streams r RTCP can synchronize different media streams within a RTP session. r Consider videoconferencing app for which each sender generates one RTP stream for video and one for audio. r Timestamps in RTP packets tied to the video and audio sampling clocks m not tied to the wall- clock time r Each RTCP sender-report packet contains (for the most recently generated packet in the associated RTP stream): m timestamp of the RTP packet m wall-clock time for when packet was created. r Receivers can use this association to synchronize the playout of audio and video.

23 Session Initiation Protocol (SIP) r Comes from IETF SIP long-term vision r All telephone calls and video conference calls take place over the Internet r People are identified by names or addresses, rather than by phone numbers r You can reach the callee, no matter where the callee roams, no matter what IP device the callee is currently using.

24 SIP Services r Setting up a call m Provides mechanisms for caller to let callee know she wants to establish a call m Provides mechanisms so that caller and callee can agree on media type and encoding. m Provides mechanisms to end call. r Determine current IP address of callee. m Maps mnemonic identifier to current IP address r Call management m Add new media streams during call m Change encoding during call m Invite others m Transfer and hold calls

25 Setting up a call to a known IP address Alice’s SIP invite message indicates her port number & IP address. Indicates encoding that Alice prefers to receive (PCM ulaw) Bob’s 200 OK message indicates his port number, IP address & preferred encoding (GSM) SIP messages can be sent over TCP or UDP; here sent over RTP/UDP. Default SIP port number is 5060.

26 Name translation and user location r Caller wants to call callee, but only has callee’s name or address. r Need to get IP address of callee’s current host: m user moves around m DHCP protocol m user has different IP devices (PC, PDA, car device) r Result can be based on: m time of day (work, home) m caller (don’t want boss to call you at home) m status of callee (calls sent to voic when callee is already talking to someone) Service provided by SIP servers: r SIP registrar server r SIP proxy server

27 SIP Registrar REGISTER sip:domain.com SIP/2.0 Via: SIP/2.0/UDP From: To: Expires: 3600 r When Bob starts SIP client, client sends SIP REGISTER message to Bob’s registrar server (similar function needed by Instant Messaging) Register Message:

28 SIP Proxy r Alice sends invite message to her proxy server m contains address r Proxy responsible for routing SIP messages to callee m possibly through multiple proxies. r Callee sends response back through the same set of proxies. r Proxy returns SIP response message to Alice m contains Bob’s IP address r Note: proxy is analogous to local DNS server

29 Example Caller with places a call to (1) Jim sends INVITE message to umass SIP proxy. (2) Proxy forwards request to upenn registrar server. (3) upenn server returns redirect response, indicating that it should try (4) umass proxy sends INVITE to eurecom registrar. (5) eurecom registrar forwards INVITE to , which is running keith’s SIP client. (6-8) SIP response sent back (9) media sent directly between clients. Note: also a SIP ack message, which is not shown.

30 Comparison with H.323 r H.323 is another signaling protocol for real-time, interactive r H.323 is a complete, vertically integrated suite of protocols for multimedia conferencing: signaling, registration, admission control, transport and codecs. r SIP is a single component. Works with RTP, but does not mandate it. Can be combined with other protocols and services. r H.323 comes from the ITU (telephony). r SIP comes from IETF: Borrows much of its concepts from HTTP. SIP has a Web flavor, whereas H.323 has a telephony flavor. r SIP uses the KISS principle: Keep it simple stupid.

31 Improving QoS in IP Networks Thus far: “making the best of best effort” Future: next generation Internet with QoS guarantees m What do we need to do to get QoS guarantees? r simple model for sharing and congestion studies:

32 Principles for QoS Guarantees r Example: 1Mbps IP phone, FTP share 1.5 Mbps link. m bursts of FTP can congest router, cause audio loss m want to give priority to audio over FTP packet marking needed for router to distinguish between different classes; and new router policy to treat packets accordingly Principle 1

33 Principles for QoS Guarantees (more) r what if applications misbehave (audio sends higher than declared rate) m policing: force source adherence to bandwidth allocations r marking and policing at network edge: m similar to ATM UNI (User Network Interface) provide protection (isolation) for one class from others Principle 2

34 Principles for QoS Guarantees (more) r Allocating fixed (non-sharable) bandwidth to flow: inefficient use of bandwidth if flows doesn’t use its allocation While providing isolation, it is desirable to use resources as efficiently as possible Principle 3

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

36 Summary of QoS Principles Let’s next look at mechanisms for achieving this ….

37 Scheduling And Policing Mechanisms r scheduling: choose next packet to send on link r FIFO (first in first out) scheduling: send in order of arrival to queue m 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

38 Scheduling Policies: more Priority scheduling: transmit highest-priority queued packet r multiple classes, with different priorities m class may depend on marking or other header info, e.g. IP source/dest, port numbers, etc..

39 Scheduling Policies: still more Weighted Fair Queuing: r generalized Round Robin r each class gets weighted amount of service in each cycle

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

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

42 Policing Mechanisms (more) r Leaky bucket + WFQ  provide guaranteed upper bound on delay, i.e., QoS guarantee! How? m WFQ: guaranteed share of bandwidth m Leaky bucket: limit max number of packets in queue (burst)

43 IETF Integrated Services (IntServ) r architecture for providing QoS guarantees in IP networks for individual application sessions r resource reservation: routers maintain state info of allocated resources, QoS req’s r admit/deny new call setup requests: Question: can newly arriving flow be admitted with performance guarantees while not violated QoS guarantees made to already admitted flows?

44 IntServ: QoS guarantee scenario r Resource reservation m call setup, signaling (RSVP) m traffic, QoS declaration m per-element admission control m QoS-sensitive scheduling (e.g., WFQ) request/ reply

45 Call Admission Arriving session must: r declare its QoS requirement m R-spec: defines the QoS being requested r characterize traffic it will send into network m T-spec: defines traffic characteristics r signaling protocol: needed to carry R-spec and T- spec to routers (where reservation is required) m RSVP

46 IntServ QoS: Service models [rfc2211, rfc 2212] Guaranteed service: r worst case traffic arrival: leaky-bucket-policed source r simple (mathematically provable) bound on delay [Parekh 1993, Cruz 1988] WFQ token rate, r bucket size, b per-flow rate, R D = b/R max arriving traffic

47 IETF Differentiated Services Concerns with IntServ: r Scalability: signaling, maintaining per-flow router state difficult with large number of flows m Example: OC-48 (2.5 Gbps) link serving 64 Kbps audio streams  39,000 flows! Each require state maintenance. r Flexible Service Models: Intserv has only two classes. Also want “qualitative” service classes m relative service distinction: Platinum, Gold, Silver DiffServ approach: r simple functions in network core, relatively complex functions at edge routers (or hosts) r Don’t define service classes, provide functional components to build service classes

48 Edge router:  per-flow traffic management  Classifies (marks) pkts  different classes  within a class: in-profile and out-profile Core router:  per class traffic management  buffering and scheduling based on marking at edge  preference given to in-profile packets DiffServ Architecture scheduling... r b marking

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

50 Edge-router: Classification and Conditioning r Packet is marked in the Type of Service (TOS) in IPv4, and Traffic Class in IPv6 r 6 bits used for Differentiated Service Code Point (DSCP) and determine Per-Hop Behavior (PHB) that the packet will receive r 2 bits are currently unused

51 Edge-router: Classification and Conditioning may be desirable to limit traffic injection rate of some class: r user declares traffic profile (e.g., rate, burst size) r traffic metered, shaped if non-conforming

52 Core-router: Forwarding (PHB) r PHB result in a different observable (measurable) forwarding performance behavior r PHB does not specify what mechanisms to use to ensure required PHB performance behavior r Examples: m Class A gets x% of outgoing link bandwidth over time intervals of a specified length m Class A packets leave first before packets from class B

53 Core-router: Forwarding (PHB) PHBs being developed: r Expedited Forwarding (EF): pkt departure rate of a class equals or exceeds specified rate m logical link with a minimum guaranteed rate m May require edge routers to limit EF traffic rate m Could be implemented using strict priority scheduling or WFQ with higher weight for EF traffic r Assured Forwarding: multiple traffic classes, treated differently m amount of bandwidth allocated, or drop priorities m Can be implemented using WFQ+leaky bucket or RED (Random Early Detection) with different threshold values. See Sections and in [PD07]

54 Multimedia Networking: Summary r multimedia applications and requirements r making the best of today’s best effort service r scheduling and policing mechanisms r next generation Internet: Intserv, RSVP, Diffserv

55 Intserv and Diffserv: Discussion r For over 20 years, many attempts failed to introduce QoS into packet-switched networks. Why? r Both Intserv and Diffserv need collaboration between all ISPs, otherwise you can not provide guarantee  hard r QoS needs accounting, extra processing (shaping, marking, policing,…)  complex and costly equipment/management  charge customers more r Most of the time there would be no perceived difference between Best-effort and Intserv/Diffserv services, at moderate load m So why would customers pay more??

56 How should the Internet evolve to better support multimedia? Integrated services philosophy: r Fundamental changes in Internet so that apps can reserve end-to-end bandwidth r Requires new, complex software in hosts & routers Laissez-faire (minimum intervention) r no major changes r more bandwidth when needed r content distribution, application-layer multicast m application layer Differentiated services philosophy: r Fewer changes to Internet infrastructure, yet provide 1st and 2nd class service. What’s your opinion?