Computer Communication Laboratory
Dept. of computing, Soonsil University
 

Quality of Service in GRID

What is GRID ?

Grid computing is a form of distributed computing that involves coordinating and sharing computing, application, data, storage, or network resources across dynamic and geographically dispersed organizations. Grid technologies promise to change the way organizations tackle complex computational problems. However, the vision of large scale resource sharing is not yet a reality in many areas ? Grid computing is an evolving area of computing, where standards and technology are still being developed to enable this new paradigm.

We participate in National GRID Construction Plan from 2002 and an area of our research is next generation Internet support to provide QoS for GRID application. Thus we develope QoS mapper for allocating bandwidth to provide more efficient environment of GRID application. The overall architecture of QoS mapper shows in Figure 1.


Figure 1. The architecture of QoS mapper

The QoS mapper uses PID feedback control system (Figure 2) which was designed from modeling of network response characteristics. The feedback mechanism executes continuous adaptive activity by changing a suitable allocation value computed form measured present throughput and previous allocated bandwidth. In the PID system, the P means proportional coefficient, the I means integration coefficient and the D means derivative coefficient. The role of proportional coefficient is executing control with proportion to the error that is the difference between target value and response value. This is the most principal control value. The role of integration coefficient is removing the steady state error of proportional control result. And the role of derivative coefficient is having the capability of prediction. We make the best use of the D to increase system response time. But if there are some noise, the system may be unstable. In this study, we construct PID controller and determine control parameters by CEMTOOL that is analytical tool for control system.


Figure 2. PID Feedback Control System

We implement QoS mapper to the real network. Our QoS mapper provide differentiated services by selecting various parameters (IP address, kinds of traffics, kinds of ToS fields, etc). These results are presented below.


Figure 3. QoS Control by ToS Field


Figure 4. QoS Control by IP Address


Figure 5. QoS Control by UDP Traffic

We expect that GRID computing and next generation Internet (IPv6) may be main infra structure of future computing environment. But in case of Korea, GRID computing is initial stage. Thus more researches for GRID network are needed.


Copyright. Computer Networking and Mobile Computing Team All right reserved.