/[gzz]/gzz/Documentation/misc/hemppah-progradu/existing_systems_overview
ViewVC logotype

Diff of /gzz/Documentation/misc/hemppah-progradu/existing_systems_overview

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 1.1 by hemppah, Wed Oct 30 11:47:30 2002 UTC revision 1.2 by hemppah, Thu Oct 31 08:15:01 2002 UTC
# Line 1  Line 1 
1    [This document is meant to be an overview of existing p2p systems.
2    Document is related to hh@gzz.info's Master Thesis project.
3    
4    Please notice that document's contents are updated constantly.]
5    
6    
7    
8    1. A summary of algorithms used in existing systems
9    
10  Comparison of efficiency of object location in existing distributed  Comparison of efficiency of object location in existing distributed
11  (p2p) systems (n is the number of nodes):  (p2p) systems (n is the number of nodes):
12    
# Line 16  How many neighbour nodes each node maint Line 25  How many neighbour nodes each node maint
25  Search:  Search:
26  Number of messages when an object lookup is performed  Number of messages when an object lookup is performed
27    
28  1. Chord  The table above has been presented in [1].
29    
30    2. A brief description of existing systems
31    
32    2.1. Chord
33    
34  -Chord constructs a distributed lookup service using a routing table of  -Chord constructs a distributed lookup service using a routing table of
35  logarithmic size.  logarithmic size.
# Line 29  maintained which cross a greater distanc Line 42  maintained which cross a greater distanc
42  -Consistent hashing (distributed hash table)  -Consistent hashing (distributed hash table)
43    
44    
45  2. CAN  2.2. CAN
46    
47  -CAN places objects into a virtual space.  -CAN places objects into a virtual space.
48  -Each peer logically occupies a zone in a virtual space (n dimensional)  -Each peer logically occupies a zone in a virtual space (n dimensional)
# Line 39  using global hash funktions Line 52  using global hash funktions
52  destination.  destination.
53  -Consistent hashing (distributed hash table)  -Consistent hashing (distributed hash table)
54    
55  3. Tapestry  2.3. Tapestry
56    
57  -Location and routing architecture (Plaxton, Rajaraman and Richa)  -Location and routing architecture (Plaxton, Rajaraman and Richa)
58  -Distributed data structure  -Distributed data structure
# Line 47  destination. Line 60  destination.
60  locality  locality
61    
62    
63  3. Gnutella  2.3. Gnutella
64    
65  -Bounded broadcast mechanism used for searching  -Bounded broadcast mechanism used for searching
66    -Power law degree distribution
67    
68  4. FreeNet  2.4. FreeNet
69    
70  -Chaotic routing scheme where objects are published to a few nearest  -Chaotic routing scheme where objects are published to a few nearest
71  neighbors  neighbors
72  -Queries follow gradients generated by object pointers  -Queries follow gradients generated by object pointers
73    -Power law degree distribution
74    
75    3. Different architectures for p2p networks [2]:
76    
77    3.1 Centralized
78    
79    -These kind of systems have a constantly updated database/directory
80    which is hosted at central locations.
81    -Nodes in the p2p networks performs a requests to the central directory
82    server to find other nodes and/or files
83    -Do not scale well
84    -Have a single point of failure
85    -Example: Napster
86    
87    3.2 Decentralized but Structured
88    
89    -Systems have no central server, but have a great amount of structure
90    -In this case, structure means that  p2p network's topology and
91    distribution of resources are tightly controlled
92    -Freenet is an example of "loosely structured" p2p system: distribution
93    of resources is based on hints
94    -CAN, Chord, Pastry, Tapestry are examples of "higly structured" p2p
95    systems: both the p2p network topology and the distribution of resources
96    are tightly controlled
97    -Hash-table-like interface
98    -Prevalant in the research literature but almost invisible on the
99    current p2p networks
100    -Isn't clear how well "highly structured" p2p systems work with in real
101    life
102    
103    
104    3.3 Decentralized and Unstructured
105    
106    -There is neither a centralized server/directory nor any precise control
107    over the network topology or resource distribution
108    -Gnutella is an example of decentralized and unstructured p2p system
109    -p2p network is is created by nodes joining the network with some loose
110    rules
111    -The resultant topology has a certain properties, but placement of
112    resources is not controlled
113    -To find a resource, node sends a request to its neighbours
114    -The most typical request method is flooding: request is propagated to
115    all neighbours within a certain radius
116    -Search mechanisms are unscalable: a great amount of loads are generated
117    on the network during search requests
118    
119    
120    Bibliography:
121    
122    [1] Hildrum Kirsten, Kubiatowicz John D., Rao Satish and Zhao Ben Y.,
123    Distributed Object Location in a Dynamic Network
124    
125    [2] Lv, Qin, Cao Pei, Cohen Edith, Li Kai, Shenker Scott, Search and
126    Replication in Unstructured Peer-to-Peer Networks

Legend:
Removed from v.1.1  
changed lines
  Added in v.1.2

savannah-hackers-public@gnu.org
ViewVC Help
Powered by ViewVC 1.1.26