基于容器云的微服务系统

doi:10.11959/j.issn.1000-0801.2018168

摘要/Abstract

摘要：

现有微服务系统持续交付仅将可执行代码或程序放入运行环境，缺乏对容器化服务的敏捷开发流程。对于容器化服务的版本测试和配置更新等相关问题也缺乏相应的解决方案。因此，如何在微服务系统上快速构建容器云并提供相关完整解决方案是急需解决的技术难题。提出了一种基于容器云的微服务系统。以容器为载体灵活快速构建微服务，为上层应用提供容器云服务。实现了服务容器化持续集成，支持多个服务的多个版本联调测试，达到生产不停机、版本无缝切换的效果；由运维监控平台对这些服务进行统一的监控管理。经过生产环境验证，系统能够保障多种服务运行不停机，短时间响应请求，满足生产的需求。

关键词: 容器云, 微服务, 持续交付, 微服务监控

Abstract:

Lack of the procedure of deploying containerized services in microservice systems,the continuous delivery only put executable codes or programs to the production environment.There is no solution for problems of the test version and configuration updates.The microservice system based on container cloud was proposed.Continuous integration for containerized services and the method that supports debugging tests for multi-version of multi-services were designed.Goals of non-stop production and seamless switch between versions were achieved.These services were monitored and managed by an operation and maintenance platform.After production environment verification,the proposed system supports services running without stop and immediate response and satisfies the needs of production.

Key words: container cloud, microservice, continuous delivery, mircoservice monitoring

杨迪. 基于容器云的微服务系统[J]. 电信科学, 2018, 34(9): 169-178.

Di YANG. Microservice system based on container cloud[J]. Telecommunications Science, 2018, 34(9): 169-178.

图/表 11

图1

图2

图3

图4

图5

图6

图7

图8

图9

表1

表2

参考文献 14

[1]	JARAMILO D , NGUYEN V , SMART R . Leveraging microservices architecture by using Docker technology[J]. SoutheastCon, 2016,5(16): 1-5.
[2]	廖建新 . 大数据技术的应用现状与展望[J]. 电信科学, 2015,31(7): 7-18. LIAO J X . Big data technology:current applications and prospects[J]. Telecommunications Science, 2015,31(7): 7-18.
[3]	陈涛, 鲁萌, 陈彦名 . 运营商大数据技术应用研究[J]. 电信科学, 2017,33(1): 130-134. CHEN T , LU M , CHEN Y M . Research on operators’ big data technologies and applications[J]. Telecommunications Science, 2017,33(1): 130-134.
[4]	汪保友, 钱晶, 袁时金 . 基于 Hadoop 的电信大数据采集方案研究与实现[J]. 电信科学, 2017,33(1): 135-142. WANG B Y , QIAN J , YUAN S J . Research and implementation on acquisition scheme of telecom big data based on Hadoop[J]. Telecommunications Science, 2017,33(1): 135-142.
[5]	YOUSIF M . Microservices[J]. IEEE Cloud Computing, 2016,3(5): 4-5.
[6]	张基, 李大中, 张呈宇 ,等. DC/OS关键技术与应用场景[J]. 电信科学, 2016,32(12): 13-19. ZHANG J , LI D Z , ZHANG C Y ,et al. Key technologies and application scenario of DC/OS[J]. Telecommunications Science, 2016,32(12): 13-19.
[7]	BALALAIE A , HEYDARNOORI A , JAMSHIDI P . Microservices architecture enables DevOps:migration to a cloud-native architecture[J]. IEEE Software, 2016,33(3): 42-52.
[8]	HINDMAN B , KONWINSKI A , ZAHARIA M ,et al. Mesos:a platform for fine-grained resource sharing in the data center[C]// The 8th USENIX conference on Networked systems design and implementation,USENIX Association,March 30-April 1,2011,Boston,USA. New York:ACM Press, 2011: 429-483.
[9]	SAHA P , GOVINDARAJU M , MARRU S ,et al. Integrating Apache Airavata with Docker,Marathon,and Mesos[J]. Concurrency and Computation:Practice and Experience, 2015(28): 1952-1959.
[10]	FIELDING R T . Architectural styles and the design of network-based software architecture[D]. Irvine:University of California,Irvine, 2000.
[11]	PARISA H , YVES L , ABDALLAH S . QoS assurance with lightvirtualization-a survey[C]// 2016 IEEE International Conference on Cloud Computing Technology and Science (CloudCom),Dec 12-15,2016,Luxembourg. Piacataway:IEEE Press, 2016: 1-12.
[12]	MERO?O-PE?UELA A , HOEKSTRA R . grlc makes GitHub taste like linked data APIs[C]// ESWC 2016 Satellite Events,May 29-June 2,2016,Heraklion,Greece.[S.l.:s.n. ], 2016: 342-353.
[13]	AHMED T M , BEZEMER C P , CHEN T H ,et al. Studying the effectiveness of application performance management (APM) tools for detecting performance regressions for web applications:an experience report[C]// The 13th International Conference on Mining Software Repositories,May 14-22,2016,Austin,USA. Piscataway:IEEE Press, 2016: 1-12.
[14]	CLINTON G , ZACHARY T . Elasticsearch:the definitive guide:a distributed real-time search and analytics engine[M]. Sebastopol: O’Reilly MediaPress, 2015.

请求数	并发量	吞吐量/(次·s^?1)	占用内存/GB	失败率
3 000	3 000	2 300	1.96/2.01	0
9 000	3 000	6 700	1.88/2.01	0
12 000	3 000	3 800	1.67/1.85	0
20 000	5 000	6 800	1.84/2.19	0
50 000	5 000	7 100	2.04/2.46	0
100 000	5 000	7 200	2.45/3.21	0
36 000	6 000	7 000	2.88/3.71	0
72 000	6 000	7 200	3.21/4.64	0
100 000	6 000	7 300	3.93/4.92	0
14 000	7 000	6 900	5.28/5.69	0
100 000	10 000	7 400	7.57/6.11	0

服务名	CPU使用率	内存使用率	入网流量/（bit·s^-1)	出网流量/(bit·s^-1)
route	30.04%	98%	345 331.11	793 472.64
sdcdr	2.53%	56.31%	1 074.76	1 265.73
blackname	13.83%	99.25%	92 141.56	208 470.4
ecs_userinfo	106.24%	97.26%	349 281.97	684 685.4