| Item type |
Trans(1) |
| 公開日 |
2022-01-05 |
| タイトル |
|
|
タイトル |
Detection of Layered Bottlenecks in Microservices |
| タイトル |
|
|
言語 |
en |
|
タイトル |
Detection of Layered Bottlenecks in Microservices |
| 言語 |
|
|
言語 |
eng |
| キーワード |
|
|
主題Scheme |
Other |
|
主題 |
[発表概要, Unrefereed Presentatin Abstract] |
| 資源タイプ |
|
|
資源タイプ識別子 |
http://purl.org/coar/resource_type/c_6501 |
|
資源タイプ |
journal article |
| 著者所属 |
|
|
|
IBM Research - Tokyo, IBM Japan, Ltd. |
| 著者所属 |
|
|
|
IBM Research - Tokyo, IBM Japan, Ltd. |
| 著者所属 |
|
|
|
IBM Research - Tokyo, IBM Japan, Ltd. |
| 著者所属 |
|
|
|
IBM Research - Tokyo, IBM Japan, Ltd. |
| 著者所属 |
|
|
|
IBM Research - Thomas J. Watson Research Center |
| 著者所属(英) |
|
|
|
en |
|
|
IBM Research - Tokyo, IBM Japan, Ltd. |
| 著者所属(英) |
|
|
|
en |
|
|
IBM Research - Tokyo, IBM Japan, Ltd. |
| 著者所属(英) |
|
|
|
en |
|
|
IBM Research - Tokyo, IBM Japan, Ltd. |
| 著者所属(英) |
|
|
|
en |
|
|
IBM Research - Tokyo, IBM Japan, Ltd. |
| 著者所属(英) |
|
|
|
en |
|
|
IBM Research - Thomas J. Watson Research Center |
| 著者名 |
Tatsushi, Inagaki
Yohei, Ueda
Moriyoshi, Ohara
Tatsuhiro, Chiba
Qi, Zhang
|
| 著者名(英) |
Tatsushi, Inagaki
Yohei, Ueda
Moriyoshi, Ohara
Tatsuhiro, Chiba
Qi, Zhang
|
| 論文抄録 |
|
|
内容記述タイプ |
Other |
|
内容記述 |
This presentation proposes a method to detect layered bottlenecks in microservices. A bottleneck in a service is a resource which limits the maximum performance of the service. Bottlenecks in software resources such as threads, locks, and channels, and hardware resources such as processors and disks, form a layered structure since a single request to the service can use multiple software resources simultaneously. Detection of such layered bottlenecks should be agile since development and operation of an application are integrated and agile. We detect layered bottlenecks in microservices by constructing a weighted dependency graph of microservices. A node represents a microservice and is weighted by the number of requests calculated from custom performance metrics. A link represents dependency between microservices which is extracted from connectivity between microservices. We detect layered bottlenecks in standard microservice benchmarks deployed on a container orchestration platform by applying the proposed method. We demonstrate that mitigating the bottlenecks improves the performance of the benchmarks. |
| 論文抄録(英) |
|
|
内容記述タイプ |
Other |
|
内容記述 |
This presentation proposes a method to detect layered bottlenecks in microservices. A bottleneck in a service is a resource which limits the maximum performance of the service. Bottlenecks in software resources such as threads, locks, and channels, and hardware resources such as processors and disks, form a layered structure since a single request to the service can use multiple software resources simultaneously. Detection of such layered bottlenecks should be agile since development and operation of an application are integrated and agile. We detect layered bottlenecks in microservices by constructing a weighted dependency graph of microservices. A node represents a microservice and is weighted by the number of requests calculated from custom performance metrics. A link represents dependency between microservices which is extracted from connectivity between microservices. We detect layered bottlenecks in standard microservice benchmarks deployed on a container orchestration platform by applying the proposed method. We demonstrate that mitigating the bottlenecks improves the performance of the benchmarks. |
| 書誌レコードID |
|
|
収録物識別子タイプ |
NCID |
|
収録物識別子 |
AA11464814 |
| 書誌情報 |
情報処理学会論文誌プログラミング（PRO）
巻 15,
号 1,
p. 14-14,
発行日 2022-01-05
|
| ISSN |
|
|
収録物識別子タイプ |
ISSN |
|
収録物識別子 |
1882-7802 |
| 出版者 |
|
|
言語 |
ja |
|
出版者 |
情報処理学会 |
IPSJ-TPRO1501008.pdf (26.7 kB)
