作者: Ignacio Bermudez , Stefano Traverso , Maurizio Munafo , Marco Mellia
DOI: 10.1109/TNSM.2014.2362357
关键词:
摘要: Clouds and CDNs are systems that tend to separate the content being requested by users from physical servers capable of serving it. From network point view, monitoring optimizing performance for traffic they generate challenging tasks, given same resource can be located in multiple places, which can, turn, change at any time. The first step understanding cloud CDN is thus engineering a platform. In this paper, we propose novel solution combines passive active measurements whose workflow has been tailored specifically characterize generated infrastructures. We validate our platform performing longitudinal characterization very well known infrastructure provider Amazon Web Services (AWS). By observing more than 50 000 Internet an Italian Service Provider, explore EC2, S3, CloudFront AWS services, unveiling their infrastructure, pervasiveness web services host, allocation policies as seen vantage points. Most importantly, observe evolution over two-year-long period. provided paper interest following: 1) developers aiming building measurement tools providers; 2) interested failure anomaly detection systems; 3) third-party service-level agreement certificators who design independently monitor performance. Finally, believe results about presented interesting among unveil properties operator view.