2023 Fiscal Year Final Research Report
Study of evaluation method for network computing performance on virtual environment
| Project/Area Number |
21K11878
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Review Section |
Basic Section 60060:Information network-related
|
|---|
| Research Institution | Daido University |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2021-04-01 – 2024-03-31
|
| Keywords | 仮想環境 / 仮想マシン / 並列分散処理 / リアルタイム処理 |
|---|
| Outline of Final Research Achievements |
To establish a model for describing VM performance, we researched a method for measuring the processing time of multiple VMs simultaneously running in a pseudo cloud environment. For the evaluation, we developed a program for enlarging still images using linear interpolation, and constructed an environment in which we could run the program simultaneously on multiple VMs and collect the processing time. Using this environment, we measured and compiled the processing time for each VM while varying the number of VMs that simultaneously executed the program from 1 to 30. As a result, we confirmed that the average processing time increases linearly due to memory contention until the number of CPU cores is exhausted, and once the CPU cores are exhausted, the processing time increases at a steeper rate due to waiting for free CPU cores, and we confirmed that VM performance can be described by a linear model.
|
| Free Research Field |
並列分散リアルタイム処理
|
| Academic Significance and Societal Importance of the Research Achievements |
本研究では，多くのICTシステムで利用されているVMの性能について，他のVMの処理がどのような影響を及ぼすのかを明らかにし，それを数式化することによって，システム構築前に予めVMの性能を予測するための技術の確立を目指している．本成果を使うことによって，複数のVMを使ったシステムを構築する際に，必要な数のVMだけを組み合わせることを可能にし，費用・エネルギー効率の良いシステムや，デッドラインが定まっているリアルタイム処理システムの設計・構築を可能にするものである．
|
