Loading...
- Type of Document: M.Sc. Thesis
- Language: Farsi
- Document No: 50069 (19)
- University: Sharif University of Technology
- Department: Computer Engineering
- Advisor(s): Hemmatyar, Ali Mohammad Afshin
- Abstract:
- The Internet was originally designed for host-to-host communications but is currently being used for data dissemination and retrieval. This structural mismatch makes the Internet an inefficient architecture. To overcome these inefficiencies, NDN has been proposed as a promising architecture over the Internet. One of the fundamental characteristics of this novel architecture is in-network caching. In this thesis, by focusing on the in-network cache management challenge of NDN, we first formally state the problem of social welfare maximization subject to the cache capacity constraint for the cache-enabled nodes, and then propose a distributed in-network caching algorithm based on matching theory to address it. The caching scheme considered by most of the proposals is on-path caching which aims to replicate objects redundantly along the delivery path. On the other hand, efficient use of in-network cache space needs co-operation among the nodes which poses a significant overhead to the network. Our proposed distributed in-network caching scheme only needs locally available information to match objects to cache-enabled nodes. We prove that the matching is stable and the outcome of the proposed algorithm is in competitive equilibrium. Further more, we show analytically that our proposed scheme converges to the centralized approach. We consider hit rate and social welfare as two metrics to evaluate the performance of our proposed caching scheme. Through extensive simulation results, convergence of our proposed scheme to the centralized solution is demonstrated and its outperformance against the RND scheme (a commonly applied caching scheme in NDN) is verified
- Keywords:
- Named Data Network (NDN) ; Cache Memory ; Matching ; Memory Management ; Cache Management Schema
- محتواي کتاب
- view
- مقدمه
- مروری بر پژوهشهای مرتبط پیشین
- بیان مسئله و روش پیشنهادی
- نتایج شبیهسازی
- جمعبندی و نتیجهگیری
- مراجع