Cryptography Reference
In-Depth Information
is boosted by 7
9%. Furthermore, penalty for having two sub-optimal entries in its
finger table is negligible and requires only a small amount of additional lookup calls
(3.4%).
The experimental results present interesting insight about the trade-offs between net-
work structure and their impact on reliability. We can deduct that additional entries in
the list have limited effect. Thus, it is better to consider other avenues of adding con-
nections in order to enhance system's reliability. Modifying the finger table can be an
option to consider because it increases coverage without deteriorating its original func-
tionality.
∼
7
Conclusions
We evaluated alert distribution systems implemented using three control channel mech-
anisms under different adversarial scenarios. Our evaluation enabled us to draw a num-
ber of interesting insights regarding the reliability of the signaling channel. The pure
distributed system (DHTs), designed to be robust under attacks, suffers in terms of net-
work performance including latency and coverage. In the case of random attack, DHTs
reliability turned out to be worse than that of a super-node based centralized design.
To alleviate this, we proposed the integration of centralized and the distributed designs.
Our approach consists of structural changes that enable us to seamlessly integrate a SN
network and a DHT network. We evaluated a hybrid network design that offered the
best coverage and reliability under all type of attack scenarios. We believe that with
proper engineering choices, we can further enhance the system's reliability.
Acknowledgements.
This work was supported by the NSF through Grant CNS-06-
27473, 09-37060 to the Computing Research Association for the CIFellows Project,by
ONR through MURI Contract N00014-07-1-0907, and by AFOSR through MURI Con-
tract FA9550-07-1-0527. Any opinions, findings, conclusions or recommendations ex-
pressed herein are those of the authors, and do not necessarily reflect those of the US
Government, ONR, AFOSR, or the NSF.
References
1. Aspnes, J., Rustagi, N., Saia, J.: Worm versus alert: Who wins in a battle for control of
a large-scale network? In: Tovar, E., Tsigas, P., Fouchal, H. (eds.) OPODIS 2007. LNCS,
vol. 4878, pp. 443-456. Springer, Heidelberg (2007)
2. Awerbuch, B., Scheideler, C.: Towards a scalable and robust dht. Theory of Computing Sys-
tems (2009)
3. Baumgart, I., Heep, B., Krause, S.: Oversim: A flexible overlay network simulation frame-
work. In: Proc. of IEEE GI (2007)
4. Bharambe, A., Herley, C., Padmanabhan, V.: Analyzing and improving a bittorrent network's
performance mechanisms. In: Proc. IEEE INFOCOM (2006)
5. Costa, M., Crowcroft, J., Castro, M., Rowstron, A., Zhou, L., Zhang, L., Barham, P.: Vigi-
lante: end-to-end containment of internet worms. In: Proc. of SOSP (2005)
6. Dabek, F., Zhao, B., Druschel, P., Kubiatowicz, J., Stoica, I.: Towards a common api for
structured peer-to-peer overlays. In: Kaashoek, M.F., Stoica, I. (eds.) IPTPS 2003. LNCS,
vol. 2735, Springer, Heidelberg (2003)
Search WWH ::
Custom Search