Simulation of ukrainian grid-infrastructure./ Shinkaryuk Dmytro Uriovich

18.01.2012

Abstract of attestation master's degree work subject:

" Simulation of ukrainian grid-infrastructure "

by Shinkaryuk Dmytro Uriovich

 

Actuality of work

Grid computing is the application of several computers to a single problem at the same time— usually to a scientific or technical problem that requires a great number of computer processing cycles or access to large amount of data. The management and scheduling of resources in such a large-scale distributed systems is complex and therefore demands sophisticated tools for analysing and fine-tuning the algorithms before applying them to the real systems. Simulation appearsto be the only feasible way to analyse algorithms on large-scale distributed systems o f heterogenous resources. Unlike using the real system in real time, simulation works well without making the analysis mechanism unnecessary complex.

 

The purpose of work

The purpose of work is simulation of ukrainian Grid-system using a GridSim toolkit, simulation of loading , investigation of ukrainian grid-infrastructure and detection of "weaknesses" of the system.

 

 Tasks solved in work

1. Creating a model of network and testing its performance under different loads should be determined: overloaded channels, site of an emerging large queues and other bottlenecks.

2. Comparison of time-shared and space-shared scheduling strategies.

 

The achieved results

Solving the tasks put in-process, an author protects:

1. Results of analysis of experiments concerning influence of different loads on computational resources and grid-network.

2. Results of analysis of comparation space-shared and time-shared modes.

3. Modification of task-grouping scheduling algorithm.

4. Results of analysis of influence of granularity time on task-grouping scheduling.

 

Scientific novelty

1. Modeled ukrainian grid-infrastructure using the latest data about configuration of computational resourses, network characteristics and processing elements evaluation.

2. Modification of task-grouping scheduling algorithm is used for modelling.

3. Efficiency of task-grouping algorithm is analyzed depending on of granularity time.

 

Practical value

Practical value of work consists in the following:

1. Experimentally investigated model of Ukrainian Grid infrastructure and tested the reaction to different loads.

2. Experimentally investigated scheduling task-grouping algorithm.

 

 Conclusions

1. Ukrainian grid-infrastructure was analyzed and modeled using the latest data about configuration of computational resourses, network characteristics and processing elements evaluation.

2. Model of a network was created and its performance was tested under different loads: overloaded channels, site of an emerging large queues and other bottlenecks.

3. Task-grouping algorithm was experimentally investigated with different values of granularity time.

4. Comparison of time-shared and space-shared scheduling strategies was done.

 

The work contains 125 pages, 43 images, 20 sources.


Keywords: GRID-INFRASTRUCTURE, TIME-SHARED AND SPACESHARED SCHEDULING, COMPUTATIONAL RESOURCES, TASKGROUPING ALGORITHM, GRIDSIM, SCHEDULER.