Visualizing QoSVisualizing QoS

Background(1/2)

•A tremendous growth in the development and deployment of networked applications such as video streaming, IP telephony, interactive games, among others.

•These applications, in contrast to elastic applications such as email and data sharing, are delay and delay jitter sensitive but can tolerate certain level of packet loss.

Background(2/2)

•Analysis of robust mechanisms for buffer management at network routers needs to be carried out in order to reduce end-to-end delay for traffic generated by various applications.

•Typical numerical experiments are included to illustrate the credibility of the proposed mechanism in the context of different quality of service (QoS) grades for various network traffic classes.

Objective

•Dynamic, generative and customizable analysis of QoS, and modify the system configuration based on the analysis:

•Dynamic and customizable analysis and visualization

•Dynamic reconfiguration of the system based on the QoS criteria

3 - layered application model• Describe system functionality, physical

device and configurations using spatial-temporal data

Process Model•A formal language of Petri-Net with a set of

entities and a set of activities scheduled according to particular sequences and logical constraints

•Use QoS analysis methods, such as static analysis and extensible logic, to see whether the information flow satisfies the requirements of performance, synchronization, service cost, and management strategy.

Programming Model

•Assign different data types to Objects and different programming languages to Processes

•Different data types and programming languages deal with different temporal and spatial tasks

Deployment Model

•Assign storage devices to data types and computing device to program codes

•Be portable to different computing platforms and physical devices

Spatial Time Data Format Design • trajectory location management• we denoted P(O)as a path which is given as a

sequence of blocks (edges) as the shortest cost (distance or travel-time) path in the map graph.

• The whole route represented by is a polyline denoted L(O).

• A c-trajectory is a sequence of straight-line,(x1,y1,t1), (x2,y2,t2)..., (xn,yn,tn) in 3- dimensional space

• Finally, the trajectory (or the uncertain trajectory) is obtained by associating an uncertainty threshold with the i'th line segment on the c-trajectory.

Spatial Time Reasoning Model

•Temporal Relationships

Spatial Time Reasoning Model

•2-Dimensional Spatial Relationships

2-dimensional temporal spaces

Visualization Based on Binding Approach•Dynamically generate layout specification

using Spatial-Temporal relationships

•Layout specifications are generated using spatial-temporal inference based on data about the system

•When the data change, the inference result will change. So it is dynamic

Visualization Based on Binding Approach• Bind layout components with different analysis

and visualization functions• Satisfy the runtime spatial and temporal

constraints• When constraints change, the function of

components will change. So it is dynamic.

Visualization Based on Binding Approach•Deploy the models onto different software

environments and physical devices

•When the binding data changes, configuration of the system will change. So it is configurable.

Internet Performance Monitoring Applications•Round-Trip Time *

▫In telecommunications, means time elapsed for a message to a remote place and back again.

▫In regards to TCP communication, the RTT time is calculated from the 3-way handshake by measuring the time between segment transmission and ACK receipt.

▫Directly affects through put rates in TCP systems.

What is included in the data• The data is about RTT of 25 websites related to

2008 Beijing Olympic Game.

• The RTT is collected from world-widely distributed clients

• Statistics are calculated, including Min, Max, Average, Middle, Quantiles & Coefficient of kurtosis.

*7 Days

A sample of the data

[1] Coefficient of kurtosis[2] Autonomous system with longest RTT[3] Autonomous system Number

Scene Graph Optimization with RRT

Reference

•Performance analysis of multimedia based web traffic with QoS constraints