Research Of User’s Privacy Protection On Cloud-based Anti-malware System

With the development of Internet, viruses and Trojans are showing explosive growth,the traditional anti-virus security anti-virus method is outdated;"cloud security" havesolved the problem of mass treatment of the virus, which can be called cloud-defensesystem. However, in the cloud-defense system, clients need to upload suspicious files forsecurity scanning, and these files may contain some personal information or evenclassified information. Transmitting this information would be an extreme violation ofprivacy because the files might contain not only personal data but also financial orclassified information. Thus, in the cloud-defense system, how to prevent the leakage ofprivacy information will be one of the most important issues.Our research seeks to protect the privacy of the candidate files uploaded by theclients in order to reduce the risk that the attackers extract sensitive data or personalinformation. We design a new system, which can reduce the risk of privacy data’s leakage,and not affect the normal anti-malware service. There are multiple-level strategies tohandle different candidate files. We analyze the office file’s binary format, and clear thedata stream in it, generate a new file which contains only micro object and some formatrelated information. We use our self-destructing data system to store general files, after thecloud finishes scanning, the files become unreadable. In the whole system, we hide theclient’s IP and MAC address to avoid the cloud to identify specific computer. Ourexperiment shows that our solution has effectively reduced the risk of privacy data’sleakage.System test results show that: Our system can provide the client basic securitydefenses, privacy protection system will not affect the normal cloud security services, andeffectively reduce the risk of privacy. Our system eliminates the privacy content in officedocuments, and the self-destructing data system ensures the scanning file be unreadablewhen the cloud anti-virus engines finish scan-ning the files. Our system has a small impactin the network’s response time, and the overall performance loss is only4.13%.