Print Email Facebook Twitter Multi-core architecture for anonymous Internet streaming Title Multi-core architecture for anonymous Internet streaming Author Stokkink, Q.A. Contributor Pouwelse, J.A. (mentor) Faculty Electrical Engineering, Mathematics and Computer Science Department Software Technology Programme Parallel and Distributed Systems group Date 2017-03-15 Abstract There are two key components for high throughput distributed anonymizing applications. The first key component is overhead due to message complexity of the utilized algorithms. The second key component is an nonscalable architecture to deal with this high throughput. These issues are compounded by the need for anonymization. Using a state of the art serialization technology has been shown to increase performance, in terms of CPU utilization, by 65%. This is due to the compression (byte stuffing) used by this technology. It also decreased lines of code in the Tribler project by roughly 2000 lines. Single-core architectures are shown to be optimizable by performing a minimum s,t-cut on the data flows within the original architecture: between the entry point and the most costly CPU-utilizing component as derived from profiling the application. This method is used on the Tribler technology to create a multi-core architecture. The resulting architecture is shown to be significantly more efficient in terms of consumed CPU for the delivered file download speed. Lastly it is shown that, by utilizing a cutting-edge cryptographic protocol, anonymizing file streaming applications can be sped up to a degree such that a web application implementations become feasible. Subject Multi-coreDistributed Systemsfile streamingpeer-to-peerTormultipathProtocol Buffersserializationmultiprocessing To reference this document use: http://resolver.tudelft.nl/uuid:3c7f869c-44b9-41f7-8deb-36ba284606cd Part of collection Student theses Document type master thesis Rights (c) 2017 Stokkink, Q.A. Files PDF MSc_Thesis_submitted.pdf 1.12 MB Close viewer /islandora/object/uuid:3c7f869c-44b9-41f7-8deb-36ba284606cd/datastream/OBJ/view