Real Time Application Simulating

RTAs are a widely used applications. There are many sub classes of applications belong to RTAs class such as voice conversational application and video streaming. Also chatting application like what we tested in the real experiments. In the simulation of RTA, we will simulate the voice and video applications. The voice application is the VoIP application. VoIP can carry text, live video, images and high quality stereo sound in addition to the screen sharing and all these depend on speed and reliability of the Internet connection. The audio stream is divided into small pieces, each small piece is small enough to t in a packet which is stamped with the destination address and sent through the network. The receiver should reconstruct the packets sequentially for ideal reproduction.

Voice over Internet protocol is a new technology that let users to make telephone calls using a broadband Internet connection instead of an analog phone line. There are many protocols used for video streaming and VoIP such as SIP, SCTP, RTP and RTCP. RTP & RTCP are working independently of the underlying transport and network layer. RTP & RTCP are a network protocol for delivering audio and video over IP networks. They are used extensively in communication and entertainment systems that involve stream- ing media such as telephony, video teleconference applications, television services and web-based push-to-talk features. RTP & RTCP run on the top of the UDP protocol.

There are many metrics which are important regards the performance of voice and video streaming applications. First metric is the QoS of the network which means that the packets of the voice or video have high priority in the network (Internet backbone), in our work we assume that this metric is veried by default. The second important metric is the data rated performance metrics such as throughput and mean packet delay. Also latency, end-to-end delay, Jitter and the packet loss rate are important metrics. Packet loss is when too much trac in the network causes the network to drop packets. Latency is the delay for packet delivery. Throughput refers to the total amount of data measured in bytes or bit per seconds. End-to-end delay is the time used by the packet to travel from node to another node. Jitter is the variations in delay of packet delivery. Finally, the user perception or the QoE metric such as R Score and the MOS. The user perception metric is depend on the opinion of RTA users. The R Score capture the eect of mouth-to-ear delay and losses in the packet-switched network. The R Score is mapped to the MOS in RTA. MOS is the opinion of RTA users about the quality levels (e.g. Good, Poor).

8.1.5.1 VoIP Simulation & Test

To simulate and test voice application, we simulate VoIP protocol in NS-2. We use CBR and VBR applications trac above to UDP and UDP sink agents in source and desti- nation of voice topology. VBR is an exponential trac which represent the other trac types on the link. The topology of testing voice applications in NS-2 is in Fig. 8.9.

We use NS-2 network link properties to introduce failures like degradation in the bandwidth of the link, the delay time on the link and the bit error rate on the link. Then we use trace les to calculate the important metrics such as the response time, latency, throughput, end to end delay, jitter, packet loss..etc. We calculate each metric one time in each case of failures cases (e.g. bandwidth degradation, increasing of delay time or BER). In case of voice application we just calculate the response time, end to end delay, jitter, packet loss, latency and throughput in the three cases of failures, the bandwidth

Figure 8.9: Voice Application Testing Topology

degradation, bit error rate and delay time increasing as a network failures. We use Perl scripting language to analysis the trace le and calculate the response time, latency, end to end delay, jitter, packet loss and network throughput.

We used the values of metrics in the three cases of failures to evaluate the perfor- mance of voice application which represent RTA. And this by plotting bandwidth (in KB/s), delay time (in milliseconds) or bit error rate value with response time, end to end delay,jitter, packet loss, latency and throughput. So, voice application performance degradation is represented by response time, end to end delay, jitter, packet loss, latency and throughput and bandwidth, delay time or BER in Fig. 8.10a, Fig. 8.10b & Fig. 8.10c.

8.1. CLOUD APPLICATION SIMULATIONS WITH FAILURES 83

(a) Metrics with Bandwidth (b) Metrics with Delay

(c) Metrics with Bit Error Rate

8.1.5.2 Video Simulation & Test

To simulate and test video stream application, we simulate RTP/RTCP protocols in NS-2. We use the RTP and RTCP agents in addition to RTP session. We construct a topology to test video application. The topology is consisting of three nodes sender, receiver and video server which connect the sender and receiver with a link. The sender and receiver nodes have the Trace Trac application above RTP agent on video sender and RTCP agent on video receiver. The topology of testing video application in NS-2 is in Fig. 8.11.

Figure 8.11: Video Application Testing Topology

We use NS-2 network link properties to introduce failures like degradation in the bandwidth of the link, the delay time on the link and the bit error rate on the link. Then we use trace les to calculate the important metrics such as the response time, latency, throughput, end to end delay, jitter, packet loss..etc. We calculate each metric one time in each case of failures cases (e.g. bandwidth degradation, increasing of delay time or BER). In case of video application, we just calculate the response time, end to end delay, jitter, packet loss, latency and throughput in the three cases of failures, the bandwidth degradation, the error bit rate and the delay time increasing as a network failures. We use Perl scripting language to analysis trace les and calculate response time, latency, end to end delay, jitter, packet loss and network throughput.

We use the values of metrics in the three cases of failures to evaluate the performance of the video streaming application which represent RTA. And this by plotting band- width (in KB/s), delay time (in milliseconds) or BER value with response time, end to end delay, jitter, packet loss, latency and throughput. So, video application performance degradation is represented by response time, end to end delay, jitter, packet loss, latency and throughput and bandwidth, delay time or bit error rate in Fig. 8.12a, Fig. 8.12b & Fig. 8.12c.

In the next couple of sections, we will discuss how we solve and mitigate the failures we introduced to application classes. We will solve all the failure by using VM migration, and specially we will solve the BER failure with FEC in addition to VM migration. The solving of failures let the performance of applications return back to acceptable state and let SLA QoS usually high, by this solving we can assure SLA services performance, QoS and response times.

8.1. CLOUD APPLICATION SIMULATIONS WITH FAILURES 85

(a) Metrics with Bandwidth (b) Metrics with Delay

(c) Metrics with Bit Error Rate