An Empirical Study of Cloud API Issues

被引：6

作者：

Li, Zhongwei ^{[1
]}

Lu, Qinghua ^{[1
]}

Zhu, Liming ^{[2
]}

Xu, Xiwei ^{[2
]}

Liu, Yue ^{[1
]}

Zhang, Weishan ^{[1
]}

机构：

[1] China Univ Petr, Coll Comp & Commun Engn, Beijing, Peoples R China

[2] CSIRO, Data61, Canberra, ACT, Australia

来源：

IEEE CLOUD COMPUTING | 2018年 / 5卷 / 02期

关键词：

API; Cloud Computing; DevOps; empirical study; fault tolerance; Reliability;

D O I：

10.1109/MCC.2018.022171668

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

With the emergence of the DevOps movement, software engineers are starting to rely on cloud platform APIs for implementing many fault tolerance, self-adaptation, and continuous delivery features such as deployment changes, scaling out/in, exception handling, backup/recovery, and migration. In this article, we present an empirical study of API issues in commercial cloud platforms. We classify the API failures and their causes, and discuss possible remedies for each category to improve the reliability of cloud applications and introduce our solutions to deal with resource characteristics faults and late timing failures.

引用

页码：58 / 72

页数：15