Quality of Service 4 Quality of service measures

Quality of Service 4 Quality of service measures the kind of service provided by the system < On systems that can offer flexible services, Qo. S allows us to compare the service received 4 MM systems consist of services 4 Examples of Multimedia Qo. S parameters: < Qo. S for Audio service: = Sample rate – 8000 samples/second = Sample resolution – 8 bits per sample < Qo. S for network service: = Throughput – 100 Mbps = Connection setup time – 50 ms Slides courtesy Prof. Nahrstedt 9/9/2020 CSE 40373/60373: Multimedia Systems page 1

Layered Model for Qo. S 9/9/2020 CSE 40373/60373: Multimedia Systems page 2

Application Qo. S Parameters 9/9/2020 CSE 40373/60373: Multimedia Systems page 3

System Qo. S Parameters 9/9/2020 CSE 40373/60373: Multimedia Systems page 4

Network Qo. S Parameters 9/9/2020 CSE 40373/60373: Multimedia Systems page 5

Qo. S Classes 4 Guaranteed Service Class < Qo. S guarantees are provided based on deterministic and statistical Qo. S parameters 4 Predictive Service Class < Qo. S parameter values are estimated and based on the past behavior of the service 4 Best Effort Service Class < There are no guarantees or only partial guarantees are provided 9/9/2020 CSE 40373/60373: Multimedia Systems page 6

Qo. S Classes (cont. ) Qo. S Class determines: (a) reliability of offered Qo. S, (b) utilization of re 9/9/2020 CSE 40373/60373: Multimedia Systems page 7

Deterministic Qo. S Parameters Single Value: Qo. S 1 – average (Qo. Save), contractual value, threshold value, target value • Pair Value: <Qo. S 1, Qo. S 2> with • < Qo. S 1 – required value; Qo. S 2 – desired value < Example: <Qo. Savg, Qo. Speak>; <Qo. Smin, Qo. Smax> 9/9/2020 CSE 40373/60373: Multimedia Systems page 8

Deterministic Qo. S Parameter Values 4 Triple of Values <Qo. S 1, Qo. S 2, Qo. S 3> < Qo. S 1 – best value < Qo. S 2 – average value < Qo. S 3 – worst value 4 Example: < <Qo. Speak, Qo. Savg, Qo. Smin>, where Qo. S is network bandwidth 9/9/2020 CSE 40373/60373: Multimedia Systems page 9

Guaranteed Qo. S 4 We need to provide 100% guarantees for Qo. S values (hard guarantees) or very close to 100% (soft guarantees) 4 Current Qo. S calculation and resource allocation are based on: < Hard upper bounds for imposed workloads < Worst case assumptions about system behavior 4 Advantages: Qo. S guarantees are satisfied even in the worst case (high reliability in guarantees) 4 Disadvantage: Over-reservation of resources, hence needless rejection of requests 9/9/2020 CSE 40373/60373: Multimedia Systems page 10

Predictive Qo. S Parameters 4 We utilize Qo. S values (Qo. S 1, . . Qo. Si) and compute average < Qo. Sbound step at K>i is Qo. SK = 1/i*∑j. Qo. Sj 4 We utilize Qo. S values (Qo. S 1, , Qo. Si) and compute maximum value < Qo. SK = max j=1, …i (Qo. Sj) 4 We utilize Qo. S values (Qo. S 1, , Qo. Si) and compute minimum value < Qo. SK = min j=1, …i (Qo. Sj) 9/9/2020 CSE 40373/60373: Multimedia Systems page 11

Best Effort Qo. S 4 No Qo. S bounds or possible very weak Qo. S bounds 4 Advantages: resource capacities can be statistically multiplexed, hence more processing requests can be granted 4 Disadvantages: Qo. S may be temporally violated 9/9/2020 CSE 40373/60373: Multimedia Systems page 12

Quality-aware Service Model 4 Quality-aware Autonomous Single Service < Consists of a set of functions < Accepts input data with Qo. S level Qo. Sin = Qo. Sin=[q 1 in, . . qnin] < Generates output data with Qo. S level Qo. Sout = Qo. Sout=[q 1 out, . . qnout] 4 Example: Video player service < Input Qo. S: [Recorded Video Frame Rate, Recorded Frame Size, Recorded Pixel Precision] < Qo. Sin=[30 fps, 640 x 480 pixels, 24 bits per pixel] < Output Qo. S: [Playback Video Frame Rate, Playback Frame Size, Playback Pixel Precision] < Qo. Sout=[20 fps, 320 x 240 pixels, 24 bits per pixel] 9/9/2020 CSE 40373/60373: Multimedia Systems page 13

Quality-aware Service Model 4 Quality-aware Composite Service < Consists of set of autonomous services that are connected into a directed acyclic graph, called service graph < Is correct if the inter-service satisfied the following relation: = Qo. Sout of Service K ‘satisfies ‘Qo. Sin of Service M iff = q. Kjout = q. Mlin for q. Mlin being single Qo. S value = q. Kjout is in q. Mlin for q. Mlin being a range of Qo. S value 4 Example: < Video-on-demand service, consists of two services: retrieval service and playback service = Output quality of the retrieval service needs to correspond to input quality of playback service, or at least falls into the range of input quality of playback service 9/9/2020 CSE 40373/60373: Multimedia Systems page 14

Relation between Qo. S and Resources 9/9/2020 CSE 40373/60373: Multimedia Systems page 15

Phase 1: Establishment Phase (Qo. S Operations) 4 Qo. S Translation at different Layers < User-Application < Application-OS/Transport Subsystem 4 Qo. S Negotiation < Negotiation of Qo. S parameters among two peers/components 4 Qo. S Routing along the end-to-end path 9/9/2020 CSE 40373/60373: Multimedia Systems page 16

Qo. S Operations within Establishment Phase User/Application Qo. S Translation Overlay P 2 P Qo. S Negotiation Application/Transp Qo. S Translation Qo. S Negotiation/ Qo. S Routing in Transport Subsyste 9/9/2020 CSE 40373/60373: Multimedia Systems page 17

Operations on Qo. S in Phase 1 (Qo. S Translations) 4 Layered Translation of Qo. S parameters (must be bidirectional) < Human (user Qo. S) – application Qo. S < Application Qo. S – system Qo. S < System Qo. S – network Qo. S 4 Media Scaling < Transparent scaling < Non-transparent scaling 9/9/2020 CSE 40373/60373: Multimedia Systems page 18

Media Scaling (Examples) 4 Audio < Transparent scaling difficult (one hears the quantization noise) < Non-transparent scaling should be used 4 Video < Temporal scaling < Spatial scaling < Color space scaling (reduction of number of entries in color space) 9/9/2020 CSE 40373/60373: Multimedia Systems page 19

Qo. S Negotiation 9/9/2020 CSE 40373/60373: Multimedia Systems page 20

Different Types of Negotiation Protocols 4 Bilateral Peer-to-Peer Negotiation < Negotiation of Qo. S parameters between equal peers in the same layer 4 Triangular Negotiation < Negotiation of Qo. S parameters between layers 4 Triangular Negotiation with Bounded Value 9/9/2020 CSE 40373/60373: Multimedia Systems page 21

Bilateral Qo. S Negotiation 9/9/2020 CSE 40373/60373: Multimedia Systems page 22

Triangular Qo. S Negotiation 9/9/2020 CSE 40373/60373: Multimedia Systems page 23

Triangular Negotiation with Bounded Value 9/9/2020 CSE 40373/60373: Multimedia Systems page 24

Qo. S Routing Node A 10 G 2 G 100 M Node B 100 M 2 G 1 G 2 G 10 G 1 G = 1 Gbps 10 G = 10 Gbps 100 M = 100 Mbps 9/9/2020 200 M End Node CSE 40373/60373: Multimedia Systems Network Route page 25

Qo. S Routing Node A 1 G If Qo. S Request on a connection from Node A to B is 150 Mbps, the Qo. S Routing question is -Does a route from A to B exist tha satisfies the Qo. S requirement? - What is the best route? Node B 10 G 2 G 100 M 200 M 2 G 2 G 10 G 1 G = 1 Gbps 10 G = 10 Gbps 100 M = 100 Mbps 9/9/2020 2 G 10 G End Node CSE 40373/60373: Multimedia Systems Network Route page 26

Qo. S Routing 4 Performed during establishment phase mostly, but also during transmission phase to adapt a route if needed 4 Need to discover route (path) that meets Qo. S requirements such as throughput, end-to-end delay, loss rate < End-to-end Throughput is a min-based metric < End-to-end Delay is additive metric 9/9/2020 CSE 40373/60373: Multimedia Systems page 27

Unicast Qo. S Routing 4 Problem Formulation: < Given a source node A, destination B, a set of Qo. S constraints C, and possibly an optimization goal, we aim to find the best feasible path from A to B which satisfies C. 4 Bandwidth-optimization problem: to find a path that has the largest bandwidth on the bottleneck link (widest path) 4 Bandwidth-constrained problem: to find a path whose bottleneck bandwidth is above a required threshold value 4 Delay-optimized problem: to find a path whose total delay is minimized 4 Delay-constrained problem: to find a path whose delay is bounded by a required value. 9/9/2020 CSE 40373/60373: Multimedia Systems page 28

Qo. S Routing Strategies 4 Source Routing < Each node maintains global state and feasible path is locally computed at the source node 4 Distributed Routing < Control messages exchanged among nodes and the state information kept at each node is collectively used for the path search 4 Hierarchical Routing < Nodes are clustered into groups creating multi-level hierarchy < One can use source routing within a cluster and distributed routing among clusters 9/9/2020 CSE 40373/60373: Multimedia Systems page 29

Multimedia Resource Management 4 Resource managers with operations and resource management protocols < Various operations must be performed by resource managers in order to provide Qo. S 4 Establishment Phase < Operations are executed where schedulable units utilizing shared resources must be admitted, reserved and allocated according to Qo. S requirements 4 Enforcement Phase < Operations are executed where reservations and allocations must be enforced, and adapted if needed 9/9/2020 CSE 40373/60373: Multimedia Systems page 30

Establishment Phase Operations 4 Qo. S to Resource Mapping < Need translation profiles 4 Resource Admission < Need admission tests to check availability of shared resources 4 Resource Reservation < Need reservation mechanisms along the end-to-end path to keep information about reservations 4 Resource Allocation 9/9/2020 CSE 40373/60373: Multimedia Systems page 31

Continuous Media Resource Model 4 One possible resource utilization model for multimedia data – Linear Bounded Arrival Process Model (LBAP) 4 LBAP models message arrival process: < M – maximum message size (in bytes) < R – maximum message rate in messages per second < B – maximum burstiness (accumulation of messages) 9/9/2020 CSE 40373/60373: Multimedia Systems page 32

LBAP Resource Model 4 If we have (M, R, B), we can predict utilization of resources: < Maximum number N of messages arriving at the resource: N = B + R x Time. Interval < Maximal Average Rate R’ (in bytes per second): R’ = M x R < Maximal Buffer Size (BS in bytes): BS = M x (B+1) 9/9/2020 CSE 40373/60373: Multimedia Systems page 33

Example of LBAP 4 Consider M = 1176 Bytes per message, R = 75 messages per second, B = 10 messages 4 During a time interval of 1 second, the maximum number of messages arriving at a resource must not exceed N = 10 messages + (75 messages/second * 1 second) = 85 messages 4 Maximum average data rate in bytes per second is R’ = 1176 bytes * 10 messages/second = 88200 bytes/second 4 Maximum buffer size in bytes in BS = 1176 bytes * (10 messages + 1) = 12936 bytes 9/9/2020 CSE 40373/60373: Multimedia Systems page 34

Admission Tests 4 Task schedulability tests for CPUs < This is done for delay guarantees 4 Packet schedulability tests for sharing host interfaces, switches < This is done for delay and jitter guarantees 4 Spatial tests for buffer allocation < This is done for delay and reliability guarantees 4 Link bandwidth tests < This is done for throughput guarantees 9/9/2020 CSE 40373/60373: Multimedia Systems page 35

Resource Reservation and Allocation 4 Two types of reservations < Pessimistic approach - Worst case reservation of resources < Optimistic approach - Average case reservation of resources 4 To implement resource reservation we need: < Resource table = to capture information about managed table (e. g. , process management PID table) < Reservation table = to capture reservation information < Reservation function = to map Qo. S to resources and operate over reservation table 9/9/2020 CSE 40373/60373: Multimedia Systems page 36

Resource Reservation 4 Two types of reservation styles: < Sender-initiated reservation < Receiver-initiated reservation 9/9/2020 CSE 40373/60373: Multimedia Systems page 37

Conclusion – Current State 4 Lack of mechanisms to support Qo. S guarantees < Need research in distributed control, monitoring, adaptation and maintenance of Qo. S mechanisms 4 Lack of overall frameworks < Need Qo. S frameworks for heterogeneous environments (diverse networks, diverse devices, diverse OS) 9/9/2020 CSE 40373/60373: Multimedia Systems page 38