The Query-Cycle Simulator for Simulating P2P Networks Mario T. Schlosser Tyson E. Condie Sepandar D. Kamvar Stanford Unive...
<ul><li>Problem:   </li></ul><ul><ul><li>Accurately Simulate Real-World P2P Networks. </li></ul></ul><ul><li>Motivation: <...
Goals <ul><li>P2P Simulator </li></ul><ul><ul><li>Descriptive </li></ul></ul><ul><ul><li>Simple </li></ul></ul><ul><ul><li...
Query Cycle Model Query Cycle 1
Query Cycle Model Query Cycle 2
Query Cycle Model Query Cycle 3
Properties to Model <ul><li>Peer Content </li></ul><ul><li>Network Parameters </li></ul><ul><li>Peer Behavior </li></ul>
Properties to Model <ul><li>Peer Content </li></ul><ul><ul><li>How Much? </li></ul></ul><ul><ul><li>What Type? </li></ul><...
Data Volume <ul><li>Observations </li></ul><ul><li>Model </li></ul>Simulator assigns # of files owned by peer i according ...
Content Type: Observations <ul><li>Content Categories </li></ul><ul><li>Zipf distribution on file popularity </li></ul>Cre...
Content Type: Model <ul><li>Modeling Content Categories: </li></ul><ul><ul><li>Assume n content categories. C={c 1 ,c 2 ,…...
Content Type: Model <ul><li>Modeling Files: </li></ul><ul><ul><li>Each distinct file f may be uniquely identified by {c,r}...
Recap on Content Assignment
Recap on Content Assignment Assign Data Volume
Recap on Content Assignment {c1, c3, c4} Assign Content Categories
Recap on Content Assignment {c1=.5, c3=.3, c4=.2} Assign Interest Level to Content Categories
Recap on Content Assignment {c1=.5, c3=.3, c4=.2} Assign Files {c,r}={c1,f1} {c,r}={c1,f7} . . .
Properties to Model <ul><li>Peer Content </li></ul><ul><li>Network Parameters </li></ul><ul><ul><li>Topology </li></ul></u...
Network Parameters <ul><li>Topology: </li></ul><ul><ul><li>Observation: Power Law Topology </li></ul></ul><ul><ul><li>Mode...
Properties to Model <ul><li>Peer Content </li></ul><ul><li>Network Parameters </li></ul><ul><li>Peer Behavior </li></ul>
Query-Cycle Model <ul><li>At each cycle, peer i may be: </li></ul><ul><ul><li>active </li></ul></ul><ul><ul><li>inactive <...
<ul><li>At each cycle, peer i may be: </li></ul><ul><ul><li>active </li></ul></ul><ul><ul><li>inactive </li></ul></ul><ul>...
<ul><li>At each cycle, peer i may be: </li></ul><ul><ul><li>active </li></ul></ul><ul><ul><li>inactive </li></ul></ul><ul>...
<ul><li>At each cycle, peer i may be: </li></ul><ul><ul><li>active </li></ul></ul><ul><ul><li>inactive </li></ul></ul><ul>...
Properties to Model <ul><li>Peer Content </li></ul><ul><li>Network Parameters </li></ul><ul><li>Peer Behavior </li></ul><u...
Uptime <ul><li>Observations </li></ul><ul><li>Model </li></ul>At each query cycle, probability of being up is drawn from d...
Queries <ul><li>Observations </li></ul><ul><ul><li>None </li></ul></ul><ul><li>Model </li></ul><ul><ul><li>Based on the id...
Responses and Downloads <ul><li>Responses </li></ul><ul><ul><li>If a peer receives a query for which it owns the file, it ...
Extensions <ul><li>Different Types of Peers </li></ul><ul><ul><li>i.e., Malicious Peers </li></ul></ul><ul><li>Different M...
Samples
Future Work <ul><li>Test predictions against observations in P2P networks “in the wild”. </li></ul><ul><li>Observations, o...
The End <ul><li>Code, demos will be available at  http://www.stanford.edu/~sdkamvar/research.html  next monday. </li></ul>
Motivation Network or peer property Affected algorithms <ul><li>Topology </li></ul><ul><li>Content distribution </li></ul>...
Query Activity <ul><li>Observations </li></ul><ul><li>Model </li></ul>At each query cycle, . . .  Saroiu,Gummandi,and Grib...
Upcoming SlideShare
Loading in...5
×

Simulator

128

Published on

Published in: Technology
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total Views
128
On Slideshare
0
From Embeds
0
Number of Embeds
0
Actions
Shares
0
Downloads
1
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Simulator

  1. 1. The Query-Cycle Simulator for Simulating P2P Networks Mario T. Schlosser Tyson E. Condie Sepandar D. Kamvar Stanford University
  2. 2. <ul><li>Problem: </li></ul><ul><ul><li>Accurately Simulate Real-World P2P Networks. </li></ul></ul><ul><li>Motivation: </li></ul><ul><ul><li>Testing P2P Algorithms. </li></ul></ul>Problem For each peer i { -Repeat until convergence { - Compute . . . - Send . . . } }
  3. 3. Goals <ul><li>P2P Simulator </li></ul><ul><ul><li>Descriptive </li></ul></ul><ul><ul><li>Simple </li></ul></ul><ul><ul><li>Easily Extensible </li></ul></ul><ul><ul><li>Make it available on the web so that people can test and compare their algorithms on a standard platform. </li></ul></ul>
  4. 4. Query Cycle Model Query Cycle 1
  5. 5. Query Cycle Model Query Cycle 2
  6. 6. Query Cycle Model Query Cycle 3
  7. 7. Properties to Model <ul><li>Peer Content </li></ul><ul><li>Network Parameters </li></ul><ul><li>Peer Behavior </li></ul>
  8. 8. Properties to Model <ul><li>Peer Content </li></ul><ul><ul><li>How Much? </li></ul></ul><ul><ul><li>What Type? </li></ul></ul><ul><li>Network Parameters </li></ul><ul><li>Peer Behavior </li></ul>
  9. 9. Data Volume <ul><li>Observations </li></ul><ul><li>Model </li></ul>Simulator assigns # of files owned by peer i according to distribution. Saroiu,Gummandi,and Gribble. A Measurement Study of Peer-to-Peer File Sharing Systems, 2002.
  10. 10. Content Type: Observations <ul><li>Content Categories </li></ul><ul><li>Zipf distribution on file popularity </li></ul>Crespo and Garcia-Molina. Semantic Overlay Networks, 2002. Korfhage, Information Storage and Retrieval, 1997. Punk Rock Hip-Hop Jazz
  11. 11. Content Type: Model <ul><li>Modeling Content Categories: </li></ul><ul><ul><li>Assume n content categories. C={c 1 ,c 2 ,…,c n } </li></ul></ul><ul><ul><li>A peer i is assigned content categories according to the Zipf distribution: </li></ul></ul><ul><ul><li>It is then assigned an interest level p(c|i) to each of the assigned content categories by a uniform random distribution. </li></ul></ul>
  12. 12. Content Type: Model <ul><li>Modeling Files: </li></ul><ul><ul><li>Each distinct file f may be uniquely identified by {c,r} </li></ul></ul><ul><ul><li>A peer is assigned files by: </li></ul></ul>
  13. 13. Recap on Content Assignment
  14. 14. Recap on Content Assignment Assign Data Volume
  15. 15. Recap on Content Assignment {c1, c3, c4} Assign Content Categories
  16. 16. Recap on Content Assignment {c1=.5, c3=.3, c4=.2} Assign Interest Level to Content Categories
  17. 17. Recap on Content Assignment {c1=.5, c3=.3, c4=.2} Assign Files {c,r}={c1,f1} {c,r}={c1,f7} . . .
  18. 18. Properties to Model <ul><li>Peer Content </li></ul><ul><li>Network Parameters </li></ul><ul><ul><li>Topology </li></ul></ul><ul><ul><li>Bandwidth </li></ul></ul><ul><li>Peer Behavior </li></ul>
  19. 19. Network Parameters <ul><li>Topology: </li></ul><ul><ul><li>Observation: Power Law Topology </li></ul></ul><ul><ul><li>Model: probability of connecting to a peer is proportional to the degree of that peer. </li></ul></ul><ul><li>Bandwidth </li></ul><ul><ul><li>Simple Bandwidth Model </li></ul></ul><ul><ul><li>Can be easily extended. </li></ul></ul>
  20. 20. Properties to Model <ul><li>Peer Content </li></ul><ul><li>Network Parameters </li></ul><ul><li>Peer Behavior </li></ul>
  21. 21. Query-Cycle Model <ul><li>At each cycle, peer i may be: </li></ul><ul><ul><li>active </li></ul></ul><ul><ul><li>inactive </li></ul></ul><ul><ul><li>or down </li></ul></ul>
  22. 22. <ul><li>At each cycle, peer i may be: </li></ul><ul><ul><li>active </li></ul></ul><ul><ul><li>inactive </li></ul></ul><ul><ul><li>or down </li></ul></ul>Query-Cycle Model <ul><li>Issues a single query. </li></ul><ul><li>Waits for incoming responses. </li></ul><ul><li>Selects a source and downloads file. </li></ul><ul><li>Also: </li></ul><ul><ul><li>Responds to queries. </li></ul></ul><ul><ul><li>Forwards query messages. </li></ul></ul>
  23. 23. <ul><li>At each cycle, peer i may be: </li></ul><ul><ul><li>active </li></ul></ul><ul><ul><li>inactive </li></ul></ul><ul><ul><li>or down </li></ul></ul>Query-Cycle Model <ul><li>Responds to queries. </li></ul><ul><li>Forwards Query Messages. </li></ul>
  24. 24. <ul><li>At each cycle, peer i may be: </li></ul><ul><ul><li>active </li></ul></ul><ul><ul><li>inactive </li></ul></ul><ul><ul><li>or down </li></ul></ul>Query-Cycle Model <ul><li>Does nothing. </li></ul>
  25. 25. Properties to Model <ul><li>Peer Content </li></ul><ul><li>Network Parameters </li></ul><ul><li>Peer Behavior </li></ul><ul><ul><li>Uptime and Session Duration </li></ul></ul><ul><ul><li>Query Activity </li></ul></ul><ul><ul><li>Queries </li></ul></ul><ul><ul><li>Query Responses </li></ul></ul><ul><ul><li>Downloads </li></ul></ul>
  26. 26. Uptime <ul><li>Observations </li></ul><ul><li>Model </li></ul>At each query cycle, probability of being up is drawn from distribution in Saroiu et al. Saroiu,Gummandi,and Gribble. A Measurement Study of Peer-to-Peer File Sharing Systems, 2002.
  27. 27. Queries <ul><li>Observations </li></ul><ul><ul><li>None </li></ul></ul><ul><li>Model </li></ul><ul><ul><li>Based on the idea that peers query for files in the same categories that they own. </li></ul></ul>
  28. 28. Responses and Downloads <ul><li>Responses </li></ul><ul><ul><li>If a peer receives a query for which it owns the file, it responds. </li></ul></ul><ul><li>Source Selection </li></ul><ul><ul><li>Random </li></ul></ul>
  29. 29. Extensions <ul><li>Different Types of Peers </li></ul><ul><ul><li>i.e., Malicious Peers </li></ul></ul><ul><li>Different Models for Different Situations </li></ul><ul><ul><li>Reputation-based source selection. </li></ul></ul><ul><ul><li>Edutella: model distribution over markups rather than content categories. </li></ul></ul><ul><ul><li>Web Services: Change models for content distribution, query activity, etc. However, parameters are the same. </li></ul></ul>
  30. 30. Samples
  31. 31. Future Work <ul><li>Test predictions against observations in P2P networks “in the wild”. </li></ul><ul><li>Observations, observations, observations. </li></ul><ul><li>Model other networks. </li></ul>
  32. 32. The End <ul><li>Code, demos will be available at http://www.stanford.edu/~sdkamvar/research.html next monday. </li></ul>
  33. 33. Motivation Network or peer property Affected algorithms <ul><li>Topology </li></ul><ul><li>Content distribution </li></ul><ul><li>Bandwidth, uptime of peers </li></ul><ul><li>Structuring algorithms </li></ul><ul><li>Whatever </li></ul><ul><li>Stability of trust algorithms </li></ul>
  34. 34. Query Activity <ul><li>Observations </li></ul><ul><li>Model </li></ul>At each query cycle, . . . Saroiu,Gummandi,and Gribble. A Measurement Study of Peer-to-Peer File Sharing Systems, 2002.
  1. A particular slide catching your eye?

    Clipping is a handy way to collect important slides you want to go back to later.

×