Multimedia on the Internet 1 T SharonA Frank

Multimedia on the Internet 1 T. Sharon-A. Frank

Is the Internet Real-Time (MM)? 2 T. Sharon-A. Frank

Internet/Multimedia Assumptions • Internet • Multimedia – – – Point-to-Point (unicast) Best-Effort Delivery Elastic Applications FIFO Packet Scheduling Provides average Packet Delay – End-to-End Reliability – Statistical Multiplexing Gain 3 T. Sharon-A. Frank – – Multipoint Soft RT Constraints Inelastic Applications Need Control over Delay and Jitter – Various Traffic Classes – Need Qo. S Guarantees

Application Taxonomy (1) Applications Elastic 4 Inelastic Elastic Applications: Inelastic Applications: Can tolerate relatively large delay variance – essentially the traditional data application. Comparatively intolerant to delay, delay variance, throughput variance and errors. T. Sharon-A. Frank

Examples of Elastic Applications • Network file service: • Email: – asynchronous – message is not real-time – delivery in several minutes is acceptable • File transfer: – interactive service – require “quick” transfer – “slow” transfer acceptable 5 T. Sharon-A. Frank – – interactive service similar to file transfer fast response required (usually over LAN) • WWW: – – interactive file access mechanism fast response required Qo. S sensitive content on WWW pages

Examples of Inelastic Applications • Real-time voice: • Streaming voice: – not interactive – end-to-end delay not important – end-to-end jitter not important – data rate and loss very important 6 – person-to-person – interactive – important to control: • end-to-end data rate • end-to-end delay • end-to-end jitter • end-to-end loss T. Sharon-A. Frank

Application Taxonomy (2) Applications Elastic Interactive Burst Best Effort Level 1 Telnet X NFS Web 7 Inelastic Interactive Asynchronous Bulk Best Effort Level 2 FTP Tolerant Intolerant Best Effort Level 3 Loose Delay Bounds Firm Delay Bounds E-Mail MM-Mail Fax Streamin g VOD Medical Imaging CAD Schemes T. Sharon-A. Frank

Qo. S Types of Service 6 Best-effort Service v r e s t n rre ols u C toc pro Qno/partial guarantees/bounds 6 Predictive Service in e ic st o m Qestimation based on past network behavior 6 Guaranteed Service Qdeterministic Qstatistical 8 T. Sharon-A. Frank

Soft RT Qo. S Guarantees • Deterministic FProvide Bounds on Performance of all Packets in a Session. • Statistical FNo more than a Specified Fraction of Packets will see Performance Below a Certain Specified Value. 9 T. Sharon-A. Frank

Deterministic RT Qo. S Guarantee • Delay: no packets delayed more than D time units on E 2 E basis (T<=D). • Loss: no packet loss occurs. • Transit Window: bound transit window (Tmax-Tmin<=E). • Queuing: the delay of every packet from session i is less than x at queue j. 10 T. Sharon-A. Frank

Statistical RT Qo. S Guarantee • Delay: no more than x% of packets have a delay larger than D (PR[T>D]<epsilon) • Loss: no more than x% of packets in a session are lost PR[Packet-loss]<epsilon • Queuing: the probability that a packet from session i has a delay greater than x is guaranteed to be less than y at queue j. 11 T. Sharon-A. Frank

Application Taxonomy (3) Applications Inelastic Elastic Interactive Burst Best Effort Level 1 Telnet X NFS Web 12 Interactive Asynchronous Bulk Best Effort Level 2 FTP Tolerant Intolerant Best Effort Level 3 Loose Delay Bounds Firm Delay Bounds E-Mail MM-Mail Fax Streamin g VOD Medical Imaging CAD Schemes Best-effort Service Grab Bandwidth No Certain Arrival Time Uses Data Immediately No Admission Control Predictive Guaranteed The Opposite T. Sharon-A. Frank Care About Average Packet Delay Quantitative Maximum Delay

Example: Playback Applications • Audio/Video Services • Soft Real-Time Tolerant Constraints sender Varying delay transmit receiver Network buffer Acquire signal, Digitize, Compress 13 Buffer, Decompress, Playback If arrives late – useless/loss. Playback point: Signal generation time + Fixed offset delay. Compute offset based on max delay: provided by network based on observed Offset delay can be adjusted T. Sharon-A. Frank delays

Internet Qo. S Models • Adaptation Model – Adapt applications • hide Internet service from the users – scaling – Adapt Internet • Differentiated Services (Diff. Serv) – simple priority • Extension Model • Integrated Services (Int. Serv) – resource reservation 14 T. Sharon-A. Frank

Adaptation Model • • Use network Feedback/Scaling Adapt applications (Scaling) Minimal changes to Internet (Diff. Serv) No need for Resource Reservation: – “Bandwidth will be infinite” When? Everywhere? Overload? – “Applications can be adaptive” Too slow? Can users adapt? – “Simple priority is sufficient” All high priority? Overload? 15 T. Sharon-A. Frank

Scaling Means to sub-sample a data stream and only present a fraction of its original content. Scaling types: "Transparent Scaling - usually by dropping some portion of the data stream. "Non-transparent Scaling usually by adjusting parameters in the coding algorithm. 16 T. Sharon-A. Frank

Scaling in Audio and Video % Audio – Transparent scaling is difficult because human ear is sensitive – usually done by changing sampling rate K Video – – – 17 Temporal scaling (drop frames) Spatial scaling (reduce resolution) Frequency scaling (reduce number of DCT coefficients) Amplitude scaling (reduce color depth) Color space scaling (reduce number of color entries or even switch to gray scale) T. Sharon-A. Frank

Audio Scaling 18 T. Sharon-A. Frank

Scaling Example: Videoconferencing 19 T. Sharon-A. Frank

Scaling Example: Videoconferencing (2) 20 T. Sharon-A. Frank

Stream Management • Managing streams is all about managing bandwidth, buffers, processing capacity and scheduling priorities – which are all needed in order to realize Qo. S guarantees. • This is not as simple as it sounds, and there’s no general agreement as to “how” it should be done. • For instance: ATM’s Qo. S (which is very “rich”) has proven to be unworkable (difficult to implement). • Another technique is the Internet’s RSVP. 21 T. Sharon-A. Frank

Improving Qo. S in IP Networks • IETF groups are working on proposals to provide better Qo. S control in IP networks, i. e. , going beyond best effort to provide some assurance for Qo. S. • Work in Progress includes Differentiated Services (Diff. Serv), RSVP and Integrated Services (Int. Serv). 22 T. Sharon-A. Frank

Differentiated Services (Diff. Serv) • Relatively simple, coarse-grained Qo. S mechanism. • Deployed in networks without needing to change the operation of the end system application. • Based around marking packets with a smallfixed bit-pattern, which maps to certain handling and forwarding criteria at each hop. 23 T. Sharon-A. Frank

Extension Model Need New Integrated Services (Int. Serv) Model? • Single Service Model – Best-effort services – Soft real-time services • Keep Internet Philosophy – – – 24 Downward compatible Common infrastructure Unified protocol stack Open/public access User usage-based pricing T. Sharon-A. Frank

Resource Reservation • Pre-allocation of needed resources to guarantee deterministic Qo. S. • Allocated resources are dedicated; if not used – remain idle. • Example: Internet RSVP – Resource re. Ser. Vation Protocol. • If resources cannot be reserved, scaling can be used. 25 T. Sharon-A. Frank

Internet RSVP Qo. S The basic organization of RSVP for resource reservation in a distributed system – transport-level control protocol for enabling resource reservations in routers. Interesting characteristic: receiver initiated. 26 T. Sharon-A. Frank

Specifying Qo. S with Flow Specifications 27 Characteristics of the Input Service Required • maximum data unit size (bytes) • Token bucket rate (bytes/sec) • Toke bucket size (bytes) • Maximum transmission rate (bytes/sec) • • • Loss sensitivity (bytes) Loss interval ( sec) Burst loss sensitivity (data units) Minimum delay noticed ( sec) Maximum delay variation ( sec) Quality of guarantee A flow specification – one way of specifying Qo. S – a little complex, but it does work (but not via a user controlled interface). T. Sharon-A. Frank

An Approach to Implementing Qo. S 28 The principle of a token bucket algorithm – a “classic” technique for controlling the flow of data (and implementing Qo. S characteristics). T. Sharon-A. Frank

Integrated Services (Int. Serv) • An architecture for providing QOS guarantees in IP networks for individual application sessions. • Relies on resource reservation. • Routers need to maintain state info, maintaining records of allocated resources and responding to new Call setup requests on that basis. 29 T. Sharon-A. Frank