assisted content delivery systems
play

assisted Content Delivery Systems Niklas Carlsson Linkping - PowerPoint PPT Presentation

Tradeoffs in Cloud and Peer- assisted Content Delivery Systems Niklas Carlsson Linkping University Gyrgy Dan KTH Royal Institute of Technology Derek Eager University of Saskatchewan Anirban Mahanti NICTA September 5, 2012 Motivation


  1. Tradeoffs in Cloud and Peer- assisted Content Delivery Systems Niklas Carlsson Linköping University György Dan KTH Royal Institute of Technology Derek Eager University of Saskatchewan Anirban Mahanti NICTA September 5, 2012

  2. Motivation  Content provider wanting to minimize its delivery cost Catalogue of many contents  Different popularity  Average service guarantees  2

  3. Motivation Cost-efficient solution must scale  with regards to both: Request rate  Number of availible contents 

  4. Motivation Cost-efficient solution must scale  with regards to both: Request rate  Number of availible contents 

  5. Motivation Cost-efficient solution must scale  with regards to both: Request rate  Number of availible contents  Chunk-based typically efficient

  6. Motivation Cost-efficient solution must scale  with regards to both: Request rate  Number of availible contents  Not as efficient in offloading server

  7. Motivation Cost-efficient solution must scale  with regards to both: Request rate  Number of availible contents 

  8. Cost model cloud peers servers  Client can download from either  Origin servers (all contents)  Cloud storage/servers (subset of contents)  Other clients (peers)

  9. Cost model cloud peers servers  Client can download from either  Origin servers (all contents)  Cloud storage/servers (subset of contents)  Other clients (peers)

  10. Cost model cloud peers servers  Client can download from either  Origin servers (all contents)  Cloud storage/servers (subset of contents)  Other clients (peers)

  11. Cost model cloud peers servers  Client can download from either  Origin servers (all contents)  Cloud storage/servers (subset of contents)  Other clients (peers)

  12. Cost model cloud peers servers  Client can download from either  Origin servers (all contents)  Cloud storage/servers (some contents)  Other clients (peers)

  13. Cost model cloud peers servers  Client can download from either  Origin servers (all contents)  Cloud storage/servers (subset of contents)  Other clients (peers)

  14. Cost model cloud peers servers Server b/w Cloud b/w  Simple cost model Cloud storage Three (3) basic cost components 

  15. Cost model cloud peers servers Server b/w Cloud b/w Cloud storage

  16. Cost model cloud peers servers Server b/w Cloud b/w Cloud storage     s c | | B c B C M i i   / i N M i M

  17. Cost model cloud peers servers Server b/w Cloud b/w Cloud storage     s c | | B c B C M i i   / i N M i M

  18. Cost model cloud peers servers Server b/w Cloud b/w Cloud storage     s c | | B c B C M i i   / i N M i M

  19. Cost model cloud peers servers Server b/w Cloud b/w Cloud storage     s c | | B c B C M i i   / i N M i M

  20. Cost model cloud peers servers Server b/w Cloud b/w Cloud storage     s c | | B c B C M i i   / i N M i M

  21. Remote access extension cloud servers peers  Extensions Full version (in paper)        s c ( 1 ) | | B c B q f C P , i i j ij ij i     i N i M j P i N i

  22. Remote access extension cloud servers peers Different remote Locations of access cost each content  Extensions Full version (in paper)        s c ( 1 ) | | B c B q f C P , i i j ij ij i     i N i M j P i N i

  23. Cost model cloud peers servers Server b/w Cloud b/w Cloud storage     s c | | B c B C M i i   / i N M i M

  24. Cost model cloud peers servers Server b/w Cloud b/w Cloud storage Minimize     s c | | B c B C M i i   / i N M i M

  25. peer seeding Bandwidth model cloud servers In-swarm  Consider single file

  26. peer seeding Bandwidth model cloud servers In-swarm  Consider single file        s c p ( ) B B B B i i i ij i  j P i

  27. peer seeding Bandwidth model cloud servers In-swarm        s c p ( ) B B B B i i i ij i  j P i

  28. peer seeding Bandwidth model cloud servers In-swarm Total required external bandwidth requirement        s c p ( ) B B B B i i i ij i  j P i

  29. peer seeding Bandwidth model cloud servers Server Cloud In-swarm Seeding Total required external bandwidth requirement        s c p ( ) B B B B i i i ij i  j P i

  30. peer seeding Bandwidth model cloud servers Server Cloud In-swarm Seeding        s c p ( ) B B B B i i i ij i  j P i

  31. peer seeding Bandwidth model cloud servers Server Cloud In-swarm Seeding        s c p ( ) B B B B i i i ij i  j P i

  32. peer seeding Bandwidth model cloud servers Server Cloud In-swarm Seeding        s c p ( ) B B B B i i i ij i  j P i

  33. peer seeding Bandwidth model cloud servers Server Cloud In-swarm Seeding        s c p ( ) B B B B i i i ij i  j P i

  34. peer seeding Bandwidth model cloud servers Server Cloud In-swarm Seeding        s c p ( ) B B B B i i i ij i  j P i

  35. peer seeding Bandwidth model cloud servers Server Cloud In-swarm Seeding Total required external bandwidth requirement        s c p ( ) B B B B i i i ij i  j P i

  36. peer seeding External bandwidth requirement cloud servers In-swarm Total required external bandwidth requirement    ( ) B i i

  37. peer seeding External bandwidth requirement cloud servers In-swarm   k ( / ) ! L U k       / L U ( ) B e L  i k ! ( 1 ) k k k

  38. External bandwidth requirement   k ( / ) ! L U k       / L U ( ) B e L  i k ! ( 1 ) k k k

  39. External bandwidth requirement Consider missing piece policy  Server upload only one piece at a time whenever there is at  least one piece missing among peer set   k ( / ) ! L U k       / L U ( ) B e L  i k ! ( 1 ) k k k

  40. External bandwidth requirement Consider missing piece policy  Server upload only one piece at a time whenever there is at  least one piece missing among peer set  Assume (for simplicity) Poisson arrivals, piece fractions, and independent pieces on  each peer   k ( / ) ! L U k       / L U ( ) B e L  i k ! ( 1 ) k k k

  41. External bandwidth requirement Consider missing piece policy  Server upload only one piece at a time whenever there is at  least one piece missing among peer set  Assume (for simplicity) Poisson arrivals, piece fractions, and independent pieces on  each peer Departure rate   k ( / ) ! L U k       / L U ( ) B e L  i k ! ( 1 ) k k k

  42. External bandwidth requirement Consider missing piece policy  Server upload only one piece at a time whenever there is at  least one piece missing among peer set  Assume (for simplicity) Poisson arrivals, piece fractions, and independent pieces on  each peer Leave k peers behind   k ( / ) ! L U k       / L U ( ) B e L  i k ! ( 1 ) k k k

  43. External bandwidth requirement Consider missing piece policy  Server upload only one piece at a time whenever there is at  least one piece missing among peer set  Assume (for simplicity) Poisson arrivals, piece fractions, and independent pieces on  each peer Peer k missing 1/(k+1) Peer k-1 missing 2/(k+1) … Peer 1 missing k/(k+1)   k ( / ) ! L U k       / L U ( ) B e L  i k ! ( 1 ) k k k

  44. External bandwidth requirement Consider missing piece policy  Server upload only one piece at a time whenever there is at  least one piece missing among peer set  Assume (for simplicity) Poisson arrivals, piece fractions, and independent pieces on  each peer Peer k missing 1/(k+1) Peer k-1 missing 2/(k+1) … Departure rate Peer 1 missing k/(k+1) Leave k peers behind   k ( / ) ! L U k       / L U ( ) B e L  i k ! ( 1 ) k k k

  45. External bandwidth requirement Consider missing piece policy  Server upload only one piece at a time whenever there is at  least one piece missing among peer set  Assume (for simplicity) Poisson arrivals, piece fractions, and independent pieces on  each peer   k ( / ) ! L U k       / L U ( ) B e L  i k ! ( 1 ) k k k

Recommend


More recommend