Implementing and operating an internet scale distributed application using service oriented architecture principles and cloud computing infrastructure
Implementing and operating an internet scale distributed application using service oriented architecture principles and cloud computing infrastructure
复制标题
使用面向服务的架构原则和云计算基础设施实施和操作互联网规模的分布式应用程序
DOI：
10.1145/1497308.1497384
复制
发表时间：
2008
期刊：
影响因子：
--
通讯作者：
Jeff Sedayao
中科院分区：
文献类型：
--
作者：
Jeff Sedayao
本文介绍了我们的经验，实现和操作一个全球分布式应用程序组成的云计算基础设施上提供的服务。英特尔需要一项服务，让我们了解来自地球仪不同地区的最终用户使用英特尔公司网站的体验。我们在PlanetLab分布式系统测试平台上构建了我们的监控系统。我们的第一个实现依赖于静态配置，需要手动干预来响应网络和节点中断。PlanetLab具有云计算和面向服务的体系结构功能，使我们能够通过使用Web服务进行虚拟机（VM）分配、VM配置、资源可用性和地理信息来提高应用程序的健壮性。我们发现，我们必须像网络设备对待网络信息一样对待服务提供的资源信息。尽管PlanetLab虚拟机和互联网连接缺乏服务级别保证，但我们构建了一个强大的全球监控服务，供英特尔网络人员使用。
This paper describes our experiences implementing and operating a globally distributed application composed of services available on a Cloud Computing infrastructure. Intel needed a service that could give us an idea of end users' experiences with Intel Corporation's website from different regions of the globe. We built our monitoring system on top of the PlanetLab distributed systems testbed. Our first implementation relied on static configurations and needed manual intervention to respond to network and node outages. PlanetLab has Cloud Computing and Service Oriented Architecture features that enabled us to increase our application's robustness by using web services for Virtual Machine (VM) allocation, VM provisioning, resource availability, and geographic information. We found that we had to treat service provided resource information like network equipment treats network information. Despite the lack of service level guarantees for PlanetLab VMs and Internet connectivity, we constructed a robust global monitoring service that is used by Intel network personnel today.