Tang, XianE2021-09-242021-09-242021-06https://hdl.handle.net/11299/224471University of Minnesota M.S. thesis. June 2021. Major: Computer Science. Advisor: HaiYang Wang. 1 computer file (PDF); vii, 64 pages.Peer-to-Peer (P2P) applications, such as BitTorrent, are generating tremendous traffic pressure on Internet Service Providers (ISPs). To better understand the elevating P2P demands, many studies explored the P2P user Interest and investigated its impact across conventional Tier-3 ISPs. These ISPs normally providing home Internet services only to a group of local users. Nowadays, more and more home Internet services providers are also content service providers(CSPs), e.g., Google and Amazon. Different from conventional ISPs, many of these ISP/CSPs provide cloud-related services to all the users around the globe. Therefore, the relationship between these special cloud-related ISPs and P2P remains largely unclear. In this paper, we take an initial yet important step towards the understanding of cloud-related ISPs. We collected the P2P user Interest data across 225 countries from 19,838 ISPs. Our measurement showed that the users in cloud-related ISPs are currently important components in P2P systems. To better understand the unique features of cloud-related ISPs, we further applied Pearson Correlation to model the similarity between different ISPs. Our analysis results indicated that the users in cloud-related ISPs have very special interests that are different from other conventional ISPs. This observation can guide the ISPs to effectively handle their traffic with caching and optimization. Moreover, it could also help us quickly identify the service type of any remote ISPs.enThesis or Dissertation