Enhancing the Quality Level Support for Real-time Multimedia - - PowerPoint PPT Presentation

Francesco Ongaro, Eduardo Cerqueira, Luca Foschini, Antonio Corradi, Mario Gerla

Department of Computer Science and Engineering, University of Bologna, Bologna, Italy Institute of Technology, Federal University of Pará, Belém, Brazil Department of Computer Science, University of California Los Angeles, Los Angeles, USA

Enhancing the Quality Level Support for Real-time Multimedia Applications in Software-Defined Networks

Mario Gerla

2

• 1. Introduction
• 2. QoS-aware Model
• 3. Implementation
• 4. Results
• 5. Conclusion

Outline

Mario Gerla

3

Introduction (I)

Motivations

ü Internet ossification phenomenon: while technology and innovation continue to evolve, our network infrastructure system has been maintained almost in the same shape for decades. ü Nowadays, Quality of Service (QoS) is a key requirement for the Internet applications that need certain QoS performance guarantees. ü Coexistence of differentiating services with different requirements in terms of bandwidth, packet loss, latency, and also various implementation technologies in networks and every day scenarios.

Mario Gerla

4

• 1. Hybrid Mininet + real H/W testbed
• 2. QoS-aware, differentiated services optimization model
• 3. Mapping of Mean Opinion Score to Bdw, delay, loss

parameters

• 4. Solution of the MCF problem using CSP solution

Contributions

Mario Gerla

5

Introduction (II)

Software-Defined Networking paradigm

SDN in conjunction with OpenFlow allows us to control the behavior of the network by decoupling the control logic and the physical layer. Specifically, the OpenFlow protocol makes the communication among the controller and the devices possible.

Mario Gerla

6

Our QoS Architecture

Mario Gerla

7

QoS aware Model

Multi-Commodity Flow and Constrained Shortest Path (MCFCSP)

Objective Function: ü cost minimization that depends on the cost computed with the delay and packet loss of the used links. Constraints: ü flow conservation, i.e. the total flow incoming into each vertex is equal to the total flow outgoing from the same vertex, with the exception of the source and the terminal.

Mario Gerla

8

QoS aware Model

Multi-Commodity Flow and Constrained Shortest Path (MCFCSP)

ü The maximum acceptable value for the packet loss Pk

max and the delay

Dk

max, imposes the limit for each service k.

ü The variables domain guarantees that the decision variable is 0 or 1. Constraints: ü The arch capacity constraint imposes a bound of the available bandwidth for each link.

Mario Gerla

9

Implementation

Hybrid Network Topology in Mininet network emulator

Mario Gerla

10

Results (I)

Short and Permanent Congestion

Time (s) Time (s) The video player is able to manage the temporary lack of bandwidth, due to a short congestion, by means of its buffer. During a permanent congestion, the video streaming throughput falls below the critical threshold

• f 2 Mbps and the video is

completely blocked.

Mario Gerla

11

Results (II)

(i) QoS aware in case of single congestion

Time (s)

Link congestion during the video streaming service. New path followed by the video streaming flow to avoid the link congestion.

The new deployed path allows us to avoid the congested link and, incidentally, keeps the v i d e o s t r e a m i n g throughput higher than the scenario without QoS.

Mario Gerla

12

Results (II)

(ii) QoS aware in case of multiple congestion

New link congestion during the video streaming service. Second new path followed by the video streaming flow to avoid the new link congestion.

Time (s) In case of multiple link c o n g e s t i o n , t h e architecture can switch to different paths to a v o i d t h e l a c k o f available bandwidth.

Mario Gerla

13

Results (III)

Multi-Commodity Flow with QoS

Link congestion during both the video streaming and the file transfer. New path followed by the video streaming and file transfer flows to avoid the link congestion.

Time (s) The QoS mathematical m o d e l t a k e s i n t o account the different s e r v i c e t y p e requirements (i.e., bandwidth, delay, and p a c k e t l o s s ) a n d assigns the best path for each flow.

Mario Gerla

14

Results (IV)

Mean Opinion Score (MOS) and Function Cost

FUNCTION COST MOS QUALITY IMPAIRMENT

< 40 5 Excellent No block at all 40 – 55 4 Good No block or a sporadic very short block 56 – 69 3 Fair A couple of short blocks (1 - 2 s) 70 – 79 2 Poor Several blocks (2 - 4 seconds long) ≥ 80 1 Bad A lot of blocks (> 10) with long duration (7 - 10 s)

Mario Gerla

15

• 1. Hybrid Mininet + real H/W testbed
• 2. QoS-aware, differentiated services optimization model
• 3. Mapping of Mean Opinion Score to Bdw, delay, loss

parameters

• 4. Solution of the MCF problem using CSP solution

We delivered what promised

Mario Gerla

16

Conclusion

ü This paper analyzes and solves the problem of managing differentiated services guaranteeing the quality level requirements for multimedia applications in SDNs. ü We achieved a better utilization of the network resources by means of the MCFCSP mathematical model in conjunction with the modular and extensible architecture proposed. ü The results show that with our QoS architecture, that continuously ascertains the network status and allocates new paths if necessary, it is possible to avoid the link congestion effects, or at least strongly reduce them. ü Incidentally, by keeping the service throughputs as high as possible, our architecture is able to enhance the quality of the service and of experience. ü We also mapped the results given by our MCFCSP model into a MOS scale to forecast an opinion score from the client point of view. ü As a future plan, we intend to exploit our architecture to manage a seamless vertical handover between different network technologies, such as Wi-Fi, WiMAX, and LTE, controlling the QoS.

1 Department of Computer Science and Engineering, University of Bologna, Bologna, Italy 2 Institute of Technology, Federal University of Pará, Belém, Brazil 3 Department of Computer Science, University of California Los Angeles, Los Angeles, USA

francesco.ongaro@studio.unibo.it 1 antonio.corradi@unibo.it 1 luca.foschini@unibo.it 1 cerqueira@ufpa.br 2 gerla@cs.ucla.edu 3