Avaliação Experimental da Escalabilidade de Sistemas P2P e um Novo Algoritmo de Controle de Taxas de Upload
Abstract
Recent studies have show that P2P systems are not always scalable, which is evidenced by a phenomenon called ”the missing piece syndrome”. This syndrome occurs when the vast majority of peers have all pieces of the file being downloaded, except one, common to all those peers. In that scenario the system’s performance is compromised. This phenomenon was detected and studied in the literature using analytical models, however, there aren’t results indicating that it really occurs in practical situations. In this context this work has two objectives. The first is to capture the occurrence of this syndrome from an experimental setup that employs a real P2P system’s protocol (BitTorrent). To our knowledge it is the first experimental study with this purpose. The second is to propose a new upload rate control algorithm to alleviate the problem. Using analytical models we study the advantages and disadvantages of the new strategy.
References
Chow, A. L. H., Golubchik, L., and Misra, V. (2008). Improving bittorrent: A simple approach. In Proceedings of the 7th International Conference on Peer-to-peer Systems, IPTPS’08, pages 8–8, Berkeley, CA, USA. USENIX Association.
de Souza e Silva, E., Le˜ao, R. M. M., Menasché, D. S., and Towsley, D. (2014a). Scalability issues in P2P systems. CoRR, abs/1405.6228.
de Souza e Silva, E., Menasché, D. S., Le˜ao, R. M., and Towsley, D. (2014b). Sobre a capacidade de serviço de sistemas p2p. In SBRC, pages 61–74.
Duerig, J., Ricci, R., Stoller, L., Strum, M., Wong, G., Carpenter, C., Fei, Z., Griffioen, J., Nasir, H., Reed, J., and Wu, X. (2012). Getting started with geni: A user tutorial. SIGCOMM Comput. Commun. Rev., 42(1):72–77.
GENI (2015). GENI Concepts. http://groups.geni.net/geni/wiki/GENIConcepts. Acessado em Fevereiro de 2015.
Hajek, B. and Zhu, J. (2010). The missing piece syndrome in peer-to-peer communication. In Information Theory Proceedings (ISIT), 2010 IEEE International Symposium on, pages 1748–1752.
Kurose, J. and Ross, K. (2013). Computer Networking A Top-Down Approach. Pearson, 6 edition.
Legout, A., Urvoy-Keller, G., and Michiardi, P. (2006). Rarest first and choke algorithms are enough. In Proceedings of the 6th ACM SIGCOMM Conference on Internet Measurement, IMC ’06, pages 203–216, New York, NY, USA. ACM.
libtorrent (2015). libtorrent. http://www.libtorrent.org/. Acessado em Fevereiro de 2015.
Menasché, D. S., de Aragão Rocha, A. A., de Souza e Silva, E., Leão, R. M. M., Towsley, D. F., and Venkataramani, A. (2010). Estimating selfsustainability in peer-to-peer swarming systems. Perform. Eval., 67(11):1243–1258.
Menasché, D. S., de Aragão Rocha, A. A., de Souza e Silva, E., Towsley, D., and Leão, R. M. M. (2011). Implications of peer selection strategies by publishers on the performance of P2P swarming systems. SIGMETRICS Performance Evaluation Review, 39(3):55–57.
Menasché, D. S., Rocha, A. A., de Souza e Silva, E., Leão, R. M., and Towsley, D. (2012). Stability of peer-to-peer swarming systems. In SBRC, pages 161–174.
opentracker (2015). opentracker. http://erdgeist.org/arts/ software/opentracker/. Acessado em Fevereiro de 2015.
Stewart,W. J. (2009). Probability, Markov Chains, Queues, and Simulation: The Mathematical Basis of Performance Modeling. Princeton University Press.
Zhu, J. and Hajek, B. (2012). Stability of a peer-to-peer communication system. IEEE Transactions on Information Theory, 58(7):4693–4713.
