1150 |
\\ \hline |
\\ \hline |
1151 |
|
|
1152 |
|
|
1153 |
\parbox{90pt}{Efficient data discovery \cite{lv02searchreplication}, \cite{osokine02distnetworks}, \cite{yang02improvingsearch}, \cite{lv02gnutellascalable}, \cite{ganesan02yappers}, \cite{adamic02localsearch}, \cite{adamic01powerlawsearch}, \cite{ripeanu02mappinggnutella}, \cite{milgram67smallworld}, \cite{adamic99small}, \cite{sterling95beowulf}, \cite{ramanathan02goodpeers}, \cite{kleinberg99small}, \cite{nips02-Kleinberg}, \cite{zhang02using}, \cite{watts00dynamics}} & |
\parbox{90pt}{Efficient and scalable data discovery \cite{lv02searchreplication}, \cite{osokine02distnetworks}, \cite{yang02improvingsearch}, \cite{lv02gnutellascalable}, \cite{ganesan02yappers}, \cite{adamic02localsearch}, \cite{adamic01powerlawsearch}, \cite{ripeanu02mappinggnutella}, \cite{milgram67smallworld}, \cite{adamic99small}, \cite{sterling95beowulf}, \cite{ramanathan02goodpeers}, \cite{kleinberg99small}, \cite{nips02-Kleinberg}, \cite{zhang02using}, \cite{watts00dynamics}} & |
1154 |
\parbox{110pt}{Find resources efficiently, if resource exists (loosely structured)} & |
\parbox{110pt}{Find resources efficiently, if resource exists (loosely structured)} & |
1155 |
\parbox{110pt}{Super nodes, node clusters, caching techiques} & |
\parbox{110pt}{Super nodes, node clusters, caching techiques} & |
1156 |
\parbox{110pt}{More efficient, less network traffic, not comparable to DHT's efficiency} |
\parbox{110pt}{More efficient, less network traffic, not comparable to DHT's efficiency} |
1206 |
\\ \hline |
\\ \hline |
1207 |
|
|
1208 |
|
|
|
\parbox{90pt}{Scalability} & |
|
|
\parbox{110pt}{Broadcasting doesn't scale when performing searches} & |
|
|
\parbox{110pt}{Super peers, peer clusters, mutual index caching} & |
|
|
\parbox{110pt}{Better scalability, but not comparable to DHTs} |
|
|
\\ \hline |
|
|
|
|
1209 |
\parbox{90pt}{Load balancing \cite{rao03loadbalancing}, \cite{ledlie02selfp2p}, \cite{byers03dhtbalancing}} & |
\parbox{90pt}{Load balancing \cite{rao03loadbalancing}, \cite{ledlie02selfp2p}, \cite{byers03dhtbalancing}} & |
1210 |
\parbox{110pt}{Random (but uniformly distributed) identifier selection could cause systen imbalance among participants with different capabilities} & |
\parbox{110pt}{Random (but uniformly distributed) identifier selection could cause systen imbalance among participants with different capabilities} & |
1211 |
\parbox{110pt}{Caching, virtual server transfers} & |
\parbox{110pt}{Caching, virtual server transfers} & |
1893 |
|
|
1894 |
\chapter{Conclusions} |
\chapter{Conclusions} |
1895 |
|
|
1896 |
|
In this thesis, we have reviewed existing Peer-to-Peer approaches, protocols and |
1897 |
|
their properties. Currently, two main Peer-to-Peer overlay approaches |
1898 |
|
exist: loosely and tightly structured ovelrays. We discussed approaches' |
1899 |
|
differences, disadvantages and advantages. |
1900 |
|
|
1901 |
|
After that, we summarized open problems in Peer-to-Peer networks. Specifically, |
1902 |
|
we divided open problems into three sub-categories: security related problems, |
1903 |
|
performance related problems and miscellaneous problems. Each of these |
1904 |
|
sub-categories have number of open problems, where there are no solutions |
1905 |
|
yet, or solutions are only partial. Much research work is required to |
1906 |
|
solve open problems. |
1907 |
|
|
1908 |
|
Then, we focused our attention to Fenfire system. First, we gave a brief |
1909 |
|
overview of Fenfire and xanalogical model. We also discussed Storm and urn-5, |
1910 |
|
which are essential parts of Fenfire's Peer-to-Peer functionality. |
1911 |
|
|
1912 |
|
In last chapter, we evaluated existing Peer-to-Peer approaches with regard |
1913 |
|
to Fenfire's needs. We proposed, that tightly structured approach is the |
1914 |
|
best alternative to our needs. First, Storm, xanalogical |
1915 |
|
model and tightly structured approach all use global unique identifiers |
1916 |
|
for identifying data. Second, our Storm design uses Semantic-Free references |
1917 |
|
for locating data in distributed networks. As the authors of \cite{balakrishnan03semanticfree}, |
1918 |
|
we also observe that tightly structured overlays provide general purpose |
1919 |
|
interface to next-generation Reference Resolution Services. Second, by using |
1920 |
|
DOLR method of tightly structured overlay, we can minimize the the lack |
1921 |
|
of locality in tightly structured overlays. Finally, we believe that issues |
1922 |
|
related to tightly structured overlays are solved in near future, because of |
1923 |
|
wide and intensive co-operation among research groups. |
1924 |
|
|
1925 |
|
|
1926 |
|
|
1927 |
\bibliographystyle{gradu} |
\bibliographystyle{gradu} |