topology management for unstructured
play

Topology Management for Unstructured Jo ao Leit ao Introduction - PowerPoint PPT Presentation

Jan 06, 2024 •277 likes •1.29k views

Topology Management for Unstructured Overlay Networks Topology Management for Unstructured Jo ao Leit ao Introduction Overlay Networks Overview CellFarm X-BOT Jo ao Leit ao Thicket OpenFire Conclusions September 5 th ,

  1. Overview Control Technique Topology Management for Unstructured Overlay Soft-constraints over neighboring relations. Networks Enforce topological properties. Jo˜ ao Leit˜ ao During the construction of the overlay. Introduction Maintain high level of randomness: Overview Overlay is flexible to deal with churn. CellFarm Keep maintenance overhead low. X-BOT Thicket Example: OpenFire Conclusions Generate dynamic cliques of nodes that can be exploited to Publications support replication and load balancing.

  2. Overview Control Technique Topology Management for Unstructured Overlay Soft-constraints over neighboring relations. Networks Enforce topological properties. Jo˜ ao Leit˜ ao During the construction of the overlay. Introduction Maintain high level of randomness: Overview Overlay is flexible to deal with churn. CellFarm Keep maintenance overhead low. X-BOT Thicket Example: OpenFire Conclusions Generate dynamic cliques of nodes that can be exploited to Publications support replication and load balancing.

  3. Overview Control Technique Topology Management for Unstructured Overlay Soft-constraints over neighboring relations. Networks Enforce topological properties. Jo˜ ao Leit˜ ao During the construction of the overlay. Introduction Maintain high level of randomness: Overview Overlay is flexible to deal with churn. CellFarm Keep maintenance overhead low. X-BOT Thicket Example: OpenFire Conclusions Generate dynamic cliques of nodes that can be exploited to Publications support replication and load balancing.

  4. Overview Bias Technique Topology Management for Unstructured Applica'on* Overlay Networks Jo˜ ao Leit˜ ao Introduction Peer.to.Peer*Service* Overview CellFarm X-BOT Overlay*Network* Thicket * OpenFire Conclusions Control % * Bias( Publications * Network*

  5. Overview Bias Technique Topology Management for Build an random overlay network. Unstructured Overlay Networks Iteratively: Jo˜ ao Leit˜ ao Swap some existing overlay links by other links that are better . Introduction Consider a particular performance criteria ( e.g. , latency). Overview Resulting overlay: CellFarm X-BOT Random in nature. Thicket Optimized to benefit the operation of a P2P service. OpenFire Conclusions Example: Publications Promote low-latency overlay links to improve the latency of an application-level broadcast service.

  6. Overview Bias Technique Topology Management for Build an random overlay network. Unstructured Overlay Networks Iteratively: Jo˜ ao Leit˜ ao Swap some existing overlay links by other links that are better . Introduction Consider a particular performance criteria ( e.g. , latency). Overview Resulting overlay: CellFarm X-BOT Random in nature. Thicket Optimized to benefit the operation of a P2P service. OpenFire Conclusions Example: Publications Promote low-latency overlay links to improve the latency of an application-level broadcast service.

  7. Overview Bias Technique Topology Management for Build an random overlay network. Unstructured Overlay Networks Iteratively: Jo˜ ao Leit˜ ao Swap some existing overlay links by other links that are better . Introduction Consider a particular performance criteria ( e.g. , latency). Overview Resulting overlay: CellFarm X-BOT Random in nature. Thicket Optimized to benefit the operation of a P2P service. OpenFire Conclusions Example: Publications Promote low-latency overlay links to improve the latency of an application-level broadcast service.

  8. Overview Embed Technique Topology Management for Unstructured Applica'on* Overlay Networks Jo˜ ao Leit˜ ao Peer.to.Peer*Service* Introduction * Overview Embed % * CellFarm * X-BOT Overlay*Network* Thicket * OpenFire Conclusions * Control % Bias( Publications * Network*

  9. Overview Embed Technique Topology Management for Considering feedback from a P2P service. Unstructured Overlay Networks Embed a secondary topology over a pure unstructured Jo˜ ao Leit˜ ao overlay network: Secondary topology can be leveraged by the service. Introduction Remaining overlay links can be used to transmit control Overview information: CellFarm X-BOT Recover secondary topology in face of failures. Thicket Fallback if secondary topology becomes compromised. OpenFire Conclusions Example Publications Embed spanning trees over an unstructured overlay to support multimedia streaming.

  10. Overview Embed Technique Topology Management for Considering feedback from a P2P service. Unstructured Overlay Networks Embed a secondary topology over a pure unstructured Jo˜ ao Leit˜ ao overlay network: Secondary topology can be leveraged by the service. Introduction Remaining overlay links can be used to transmit control Overview information: CellFarm X-BOT Recover secondary topology in face of failures. Thicket Fallback if secondary topology becomes compromised. OpenFire Conclusions Example Publications Embed spanning trees over an unstructured overlay to support multimedia streaming.

  11. Overview Embed Technique Topology Management for Considering feedback from a P2P service. Unstructured Overlay Networks Embed a secondary topology over a pure unstructured Jo˜ ao Leit˜ ao overlay network: Secondary topology can be leveraged by the service. Introduction Remaining overlay links can be used to transmit control Overview information: CellFarm X-BOT Recover secondary topology in face of failures. Thicket Fallback if secondary topology becomes compromised. OpenFire Conclusions Example Publications Embed spanning trees over an unstructured overlay to support multimedia streaming.

  12. Overview Enrich Technique Topology Management for Unstructured Applica'on* Overlay Networks Jo˜ ao Leit˜ ao Peer.to.Peer*Service* Introduction * Overview Embed % * Enrich % CellFarm * X-BOT Overlay*Network* Thicket * OpenFire Conclusions * Control % Bias( Publications * Network*

  13. Overview Enrich Technique Topology Management for Unstructured Overlay Considering feedback from a P2P service. Networks Jo˜ ao Leit˜ ao Create temporary overlay links at the P2P service layer: Temporary links are outside the logic of the unstructured Introduction overlay. Overview Enable the service to contact peers that could not be CellFarm available through the overlay. X-BOT Thicket Example OpenFire Conclusions Create temporary overlay links that enable one to circumvent Publications Firewalls and NAT boxes.

  14. Overview Topology Management for Unstructured Applica'on* Overlay Networks Jo˜ ao Leit˜ ao Peer.to.Peer*Service* Introduction * Overview Embed % * Enrich % CellFarm * X-BOT Overlay*Network* Thicket * OpenFire Conclusions * Control % Bias( Publications * Network*

  15. Overview CellFarm Topology Management for Unstructured Applica'on* Overlay Networks Jo˜ ao Leit˜ ao Peer.to.Peer*Service* Introduction * Overview Embed % * Enrich % CellFarm * X-BOT Overlay*Network* Thicket * OpenFire Control:( * Conclusions Bias( CellFarm% Publications * Network*

  16. Overview X-BOT Topology Management for Unstructured Applica'on* Overlay Networks Jo˜ ao Leit˜ ao Peer.to.Peer*Service* Introduction * Overview Embed % * Enrich % CellFarm * X-BOT Overlay*Network* Thicket * OpenFire Control:( Bias:( * Conclusions CellFarm% X/BOT% * Publications Network*

  17. Overview Thicket Topology Management for Unstructured Applica'on* Overlay Networks Jo˜ ao Leit˜ ao Peer.to.Peer*Service* Introduction * Embed:( Overview Enrich( * Thicket% CellFarm * X-BOT Overlay*Network* Thicket * OpenFire Control:( Bias:( * Conclusions CellFarm% X/BOT% Publications * Network*

  18. Overview OpenFire Topology Management for Unstructured Applica'on* Overlay Networks Jo˜ ao Leit˜ ao Peer.to.Peer*Service* Introduction * Embed:( Enrich:( Overview * Thicket( OpenFire( CellFarm * X-BOT Overlay*Network* Thicket * OpenFire Control:( Bias:( Conclusions * CellFarm( X/BOT( Publications * Network*

  19. Overview Topology Management for Unstructured Overlay Networks Jo˜ ao Leit˜ ao Introduction Very briefly motivate the problem and solution. Overview CellFarm Present some relevant results. X-BOT Thicket OpenFire Conclusions Publications

  20. Overview Topology Management for Unstructured Overlay Networks Jo˜ ao Leit˜ ao Introduction Very briefly motivate the problem and solution. Overview CellFarm Present some relevant results. X-BOT Thicket OpenFire Conclusions Publications

  21. Control the Topology Topology Management for Unstructured Applica'on* Overlay Networks Jo˜ ao Leit˜ ao Peer.to.Peer*Service* Introduction * Embed:( Enrich:( Overview * Thicket( OpenFire( CellFarm * X-BOT Overlay*Network* Thicket * OpenFire Control:( Bias:( Conclusions * CellFarm( X/BOT( Publications * Network*

  22. CellFarm Topology Management for Unstructured Overlay Networks Jo˜ ao Leit˜ ao Goal: Offer support to replication and load balancing: Introduction Process groups. Overview This can be achieved by manipulating the topology: CellFarm Nodes self-organize in (fully connected) cliques . X-BOT Cliques are used as a replication and load balancing unit. Thicket Cliques are highly connected among them. OpenFire Conclusions Publications

  23. CellFarm Topology Management for Unstructured Overlay Networks Jo˜ ao Leit˜ ao Goal: Offer support to replication and load balancing: Introduction Process groups. Overview This can be achieved by manipulating the topology: CellFarm Nodes self-organize in (fully connected) cliques . X-BOT Cliques are used as a replication and load balancing unit. Thicket Cliques are highly connected among them. OpenFire Conclusions Publications

  24. CellFarm Topology Management for Unstructured Overlay Networks Jo˜ ao Leit˜ ao Employed technique: Control . Introduction Flexibility is essential: Overview CellFarm Enable the overlay to cope with churn. X-BOT Clique size governed by three parameters: target size; minimum size; maximum size; Thicket OpenFire Conclusions Publications

  25. CellFarm Topology Management for Unstructured Overlay Networks Jo˜ ao Leit˜ ao Employed technique: Control . Introduction Flexibility is essential: Overview CellFarm Enable the overlay to cope with churn. X-BOT Clique size governed by three parameters: target size; minimum size; maximum size; Thicket OpenFire Conclusions Publications

  26. CellFarm Main Results: Cell Size Distribution Topology Management for 6" Unstructured Overlay Networks 5" Jo˜ ao Leit˜ ao 4" Number'of'Cells' Introduction Overview 3" CellFarm X-BOT 2" Thicket OpenFire 1" Conclusions Publications 0" 8" 9" 10" 11" 12" 13" 14" 15" 16" Cell'Size' '

  27. CellFarm Main Results: Robustness to Churn Topology Management 100" for Unstructured Overlay 90" Number"of"nodes"in"largest"connected" Networks CellFarm( 80" Jo˜ ao Leit˜ ao Unstrctured( 70" Introduction compoenent"(%) " 60" Overview 50" CellFarm 40" X-BOT Thicket 30" OpenFire 20" Conclusions 10" Publications 0" 0.1" 0.2" 0.3" 0.4" 0.5" 1" 2" 3" 4" 5" Churn"rate"(%)"

  28. CellFarm Summary Topology Management for Unstructured Overlay Networks Jo˜ ao Leit˜ ao CellFarm, based on the control technique operating at the overlay network layer. Introduction Overview Control technique allows to build and maintain robust and CellFarm flexible overlays with pre-determined topological X-BOT properties. Thicket Benefit the operation of P2P resource location systems OpenFire based on unstructured overlays and one-hop replication. Conclusions Publications

  29. CellFarm Summary Topology Management for Unstructured Overlay Networks Jo˜ ao Leit˜ ao CellFarm, based on the control technique operating at the overlay network layer. Introduction Overview Control technique allows to build and maintain robust and CellFarm flexible overlays with pre-determined topological X-BOT properties. Thicket Benefit the operation of P2P resource location systems OpenFire based on unstructured overlays and one-hop replication. Conclusions Publications

  30. Overview Topology Management for Unstructured Applica'on* Overlay Networks Jo˜ ao Leit˜ ao Peer.to.Peer*Service* Introduction * Embed:( Enrich:( Overview * Thicket( OpenFire( CellFarm * X-BOT Overlay*Network* Thicket * OpenFire Control:( Bias:( Conclusions * CellFarm( X/BOT( Publications * Network*

  31. Bias the Topology Topology Management for Unstructured Applica'on* Overlay Networks Jo˜ ao Leit˜ ao Peer.to.Peer*Service* Introduction * Embed:( Enrich:( Overview * Thicket( OpenFire( CellFarm * X-BOT Overlay*Network* Thicket * OpenFire Control:( Bias:( Conclusions * CellFarm( X/BOT( Publications * Network*

  32. X-BOT Topology Management for Unstructured Overlay Networks Goal: Deal with the topology mismatch problem: Jo˜ ao Leit˜ ao Random topology means that many links may be sub-optimal. Introduction Overview This can be achieved by manipulating the topology: CellFarm Cannot employ control technique. X-BOT New nodes may offer new opportunities to improve the Thicket overlay topology. OpenFire Allow nodes to replace sub-optimal links by better links. Approach can be used to improve the overlay for different Conclusions criteria. Publications

  33. X-BOT Topology Management for Unstructured Overlay Networks Goal: Deal with the topology mismatch problem: Jo˜ ao Leit˜ ao Random topology means that many links may be sub-optimal. Introduction Overview This can be achieved by manipulating the topology: CellFarm Cannot employ control technique. X-BOT New nodes may offer new opportunities to improve the Thicket overlay topology. OpenFire Allow nodes to replace sub-optimal links by better links. Approach can be used to improve the overlay for different Conclusions criteria. Publications

  34. X-BOT Topology Management for Unstructured Overlay Employed technique: Bias . Networks Jo˜ ao Leit˜ ao Generic approach: Companion oracle tags overlay links with costs . Introduction Overview X-BOT strives to minimize the overlay cost. CellFarm Some minimal coordination is essential: X-BOT Protect relevant properties of an unstructured overlay Thicket ( e.g. , connectivity). OpenFire Rely on 4-node coordination technique to bias the overlay. Conclusions X-BOT also includes other mechanisms to protect relevant Publications overlay properties.

  35. X-BOT Topology Management for Unstructured Overlay Employed technique: Bias . Networks Jo˜ ao Leit˜ ao Generic approach: Companion oracle tags overlay links with costs . Introduction Overview X-BOT strives to minimize the overlay cost. CellFarm Some minimal coordination is essential: X-BOT Protect relevant properties of an unstructured overlay Thicket ( e.g. , connectivity). OpenFire Rely on 4-node coordination technique to bias the overlay. Conclusions X-BOT also includes other mechanisms to protect relevant Publications overlay properties.

  36. X-BOT Topology Management for Unstructured Overlay Employed technique: Bias . Networks Jo˜ ao Leit˜ ao Generic approach: Companion oracle tags overlay links with costs . Introduction Overview X-BOT strives to minimize the overlay cost. CellFarm Some minimal coordination is essential: X-BOT Protect relevant properties of an unstructured overlay Thicket ( e.g. , connectivity). OpenFire Rely on 4-node coordination technique to bias the overlay. Conclusions X-BOT also includes other mechanisms to protect relevant Publications overlay properties.

  37. X-BOT Main Results: Overlay Cost Topology Management for T-Man Unstructured 2 x 10 8 Overlay Araneola Networks GoCast X-BOT Jo˜ ao Leit˜ ao 1.5 x 10 8 Introduction overlay cost Overview CellFarm 1 x 10 8 X-BOT Thicket 5 x 10 7 OpenFire Conclusions Publications 0 0 50 100 150 200 250 cycle

  38. X-BOT Main Results: Broadcast Latency & Reliability Topology Management for Unstructured Overlay Networks Jo˜ ao Leit˜ ao Gossip-based Broadcast Introduction Latency (ms) Reliability (%) Overview Araneola 3517 . 0 100 . 00000 CellFarm GoCast 2108 . 00 99 , 99996 X-BOT T-Man 2545 . 2 13 . 80600 Thicket X-BOT 1879 . 8 100 . 00000 OpenFire Conclusions Publications

  39. X-BOT Main Results: Broadcast Latency Topology Management for Unstructured Overlay Networks Jo˜ ao Leit˜ ao Inet-3.0 Scenario Introduction Latency (ms) Reliability (%) Overview Araneola 3517 . 0 100 . 00000 CellFarm GoCast 2108 . 00 99 , 99996 X-BOT T-Man 2545 . 2 13 . 80600 Thicket X-BOT 1879 . 8 100 . 00000 OpenFire Conclusions Publications

  40. X-BOT Summary Topology Management for Unstructured Overlay Networks X-BOT, based on the bias technique operating at the Jo˜ ao Leit˜ ao overlay network layer. Introduction Bias technique allows to achieve improved overlay Overview topologies that can better match the requirements of CellFarm services executing on top of them. X-BOT Thicket If a minimal amount of coordination exists among nodes, OpenFire biasing can be achieved without compromising relevant Conclusions overlay properties. Publications

  41. X-BOT Summary Topology Management for Unstructured Overlay Networks X-BOT, based on the bias technique operating at the Jo˜ ao Leit˜ ao overlay network layer. Introduction Bias technique allows to achieve improved overlay Overview topologies that can better match the requirements of CellFarm services executing on top of them. X-BOT Thicket If a minimal amount of coordination exists among nodes, OpenFire biasing can be achieved without compromising relevant Conclusions overlay properties. Publications

  42. Bias the Topology Topology Management for Unstructured Applica'on* Overlay Networks Jo˜ ao Leit˜ ao Peer.to.Peer*Service* Introduction * Embed:( Enrich:( Overview * Thicket( OpenFire( CellFarm * X-BOT Overlay*Network* Thicket * OpenFire Control:( Bias:( Conclusions * CellFarm( X/BOT( Publications * Network*

  43. Embed the Topology Topology Management for Unstructured Applica'on* Overlay Networks Jo˜ ao Leit˜ ao Peer.to.Peer*Service* Introduction * Embed:( Enrich:( Overview * Thicket( OpenFire( CellFarm * X-BOT Overlay*Network* Thicket * OpenFire Control:( Bias:( Conclusions * CellFarm( X/BOT( Publications * Network*

  44. Thicket Topology Management for Unstructured Overlay Goal: Support efficient and robust data streaming: Networks Efficient: Rely on spanning tree. Jo˜ ao Leit˜ ao Robust: Multiple spanning trees. Introduction Load distribution: Multiple interior-node disjoint trees. Overview Efficient spanning tree topologies should consider feedback CellFarm of streaming service: X-BOT Cannot employ a technique at the overlay layer. Thicket Building independent trees is not an option. OpenFire Use a single unstructured overlay networks as a base. Conclusions Embed several spanning trees over that overlay in a Publications coordinated fashion.

  45. Thicket Topology Management for Unstructured Overlay Goal: Support efficient and robust data streaming: Networks Efficient: Rely on spanning tree. Jo˜ ao Leit˜ ao Robust: Multiple spanning trees. Introduction Load distribution: Multiple interior-node disjoint trees. Overview Efficient spanning tree topologies should consider feedback CellFarm of streaming service: X-BOT Cannot employ a technique at the overlay layer. Thicket Building independent trees is not an option. OpenFire Use a single unstructured overlay networks as a base. Conclusions Embed several spanning trees over that overlay in a Publications coordinated fashion.

  46. Thicket Topology Management for Unstructured Overlay Employed technique: Embed . Networks Coordination is essential. Jo˜ ao Leit˜ ao Overlay links used to embed a tree should consider the Introduction existence of other trees. Overview Trees should be able to recover from partitions efficiently. CellFarm The maximum forwarding load imposed over nodes should X-BOT be limited. Thicket Multiples trees allow to transmit redundant data: OpenFire e.g. , Network coding techniques. Conclusions Sustain the data stream even if one tree becomes Publications temporarily partitioned.

  47. Thicket Topology Management for Unstructured Overlay Employed technique: Embed . Networks Coordination is essential. Jo˜ ao Leit˜ ao Overlay links used to embed a tree should consider the Introduction existence of other trees. Overview Trees should be able to recover from partitions efficiently. CellFarm The maximum forwarding load imposed over nodes should X-BOT be limited. Thicket Multiples trees allow to transmit redundant data: OpenFire e.g. , Network coding techniques. Conclusions Sustain the data stream even if one tree becomes Publications temporarily partitioned.

  48. Thicket Topology Management for Unstructured Overlay Employed technique: Embed . Networks Coordination is essential. Jo˜ ao Leit˜ ao Overlay links used to embed a tree should consider the Introduction existence of other trees. Overview Trees should be able to recover from partitions efficiently. CellFarm The maximum forwarding load imposed over nodes should X-BOT be limited. Thicket Multiples trees allow to transmit redundant data: OpenFire e.g. , Network coding techniques. Conclusions Sustain the data stream even if one tree becomes Publications temporarily partitioned.

  49. Thicket Main Results: K-interior node distribution (PlanetLab) Topology Management for Thicket Unstructured 100 number of interior nodes (%) Overlay Networks Jo˜ ao Leit˜ ao 80 Introduction Overview 60 CellFarm X-BOT Thicket 40 OpenFire Conclusions 20 Publications 0 0 1 2 3 4 5 number of trees

  50. Thicket Main Results: Streaming service robustness in face of targeted node failures Topology Management for NUTS Unstructured 105 BOLTS Overlay Thicket Networks 100 Jo˜ ao Leit˜ ao 95 Introduction reliability (%) 90 Overview CellFarm 85 X-BOT 80 Thicket 75 OpenFire Conclusions 70 Publications 65 60 0 50 100 150 200 simulation cycle

  51. Thicket Summary Topology Management for Unstructured Overlay Networks Thicket, based on the embed technique operating at the Jo˜ ao Leit˜ ao peer-to-peer service layer. Introduction Embed technique allows to efficiently embed interior-node Overview disjoint trees over a single unstructured overlay. CellFarm Combines the best of gossip-based and tree-based X-BOT dissemination approaches. Thicket OpenFire Improved the usage of available resources when compared Conclusions with existing solutions. Publications

  52. Thicket Summary Topology Management for Unstructured Overlay Networks Thicket, based on the embed technique operating at the Jo˜ ao Leit˜ ao peer-to-peer service layer. Introduction Embed technique allows to efficiently embed interior-node Overview disjoint trees over a single unstructured overlay. CellFarm Combines the best of gossip-based and tree-based X-BOT dissemination approaches. Thicket OpenFire Improved the usage of available resources when compared Conclusions with existing solutions. Publications

  53. Embed the Topology Topology Management for Unstructured Applica'on* Overlay Networks Jo˜ ao Leit˜ ao Peer.to.Peer*Service* Introduction * Embed:( Enrich:( Overview * Thicket( OpenFire( CellFarm * X-BOT Overlay*Network* Thicket * OpenFire Control:( Bias:( Conclusions * CellFarm( X/BOT( Publications * Network*

  54. Enrich the Topology Topology Management for Unstructured Applica'on* Overlay Networks Jo˜ ao Leit˜ ao Peer.to.Peer*Service* Introduction * Embed:( Enrich:( Overview * Thicket( OpenFire( CellFarm * X-BOT Overlay*Network* Thicket * OpenFire Control:( Bias:( * Conclusions CellFarm( X/BOT( Publications * Network*

  55. OpenFire Topology Management for Unstructured Overlay Goal: Deal with the presence of Firewalls and NAT boxes. Networks Firewalls and NAT boxes make nodes unreachable. Jo˜ ao Leit˜ ao This can produce an unbalanced behavior in P2P services Introduction ( e.g , rumor mongering protocols). Overview Most nodes interact only with public nodes in the Internet. CellFarm Those nodes can easily became resource exhausted. X-BOT Circumvent Firewalls and NAT boxes: Thicket Taking into consideration communication patters at the OpenFire P2P service layer. Conclusions Keep connections established by confined nodes open to be Publications used them later.

  56. OpenFire Topology Management for Unstructured Overlay Goal: Deal with the presence of Firewalls and NAT boxes. Networks Firewalls and NAT boxes make nodes unreachable. Jo˜ ao Leit˜ ao This can produce an unbalanced behavior in P2P services Introduction ( e.g , rumor mongering protocols). Overview Most nodes interact only with public nodes in the Internet. CellFarm Those nodes can easily became resource exhausted. X-BOT Circumvent Firewalls and NAT boxes: Thicket Taking into consideration communication patters at the OpenFire P2P service layer. Conclusions Keep connections established by confined nodes open to be Publications used them later.

  57. OpenFire Topology Management for Unstructured Employed technique: Enrich . Overlay Networks Use a single temporary overlay link. Jo˜ ao Leit˜ ao Whenever a node receives a request from a peer it updates Introduction its temporary overlay link. Overview Nodes only process a (rumor mongering) request for each CellFarm request they sent themselves. X-BOT Additional received requests are forwarded across the temporary connection. Thicket OpenFire This solution is interesting: Conclusions Can easily ensure balanced processing of messages. Publications Nodes are not required to be aware of Firewalls and NATs explicitly.

  58. OpenFire Topology Management for Unstructured Employed technique: Enrich . Overlay Networks Use a single temporary overlay link. Jo˜ ao Leit˜ ao Whenever a node receives a request from a peer it updates Introduction its temporary overlay link. Overview Nodes only process a (rumor mongering) request for each CellFarm request they sent themselves. X-BOT Additional received requests are forwarded across the temporary connection. Thicket OpenFire This solution is interesting: Conclusions Can easily ensure balanced processing of messages. Publications Nodes are not required to be aware of Firewalls and NATs explicitly.

  59. OpenFire Topology Management for Unstructured Employed technique: Enrich . Overlay Networks Use a single temporary overlay link. Jo˜ ao Leit˜ ao Whenever a node receives a request from a peer it updates Introduction its temporary overlay link. Overview Nodes only process a (rumor mongering) request for each CellFarm request they sent themselves. X-BOT Additional received requests are forwarded across the temporary connection. Thicket OpenFire This solution is interesting: Conclusions Can easily ensure balanced processing of messages. Publications Nodes are not required to be aware of Firewalls and NATs explicitly.

  60. OpenFire Main Results: Maximum gossip exchanges per node Topology Management for **+","!" Unstructured **+",")" Overlay **+","-" Networks **+","!$" Jo˜ ao Leit˜ ao !"#$%&'()'*(++,-'%./01!*%+' Introduction Overview CellFarm X-BOT Thicket OpenFire Conclusions Publications !" !#$$" %!$$" &#$$" #!$$" '#$$" (!$$" !$#$$" !)!$$" !"#$%&'()'/(!2!%#%!3'4(#1,!+'

  61. OpenFire Main Results: Maximum latency Topology Management for '&!" ++,"-"'" Unstructured ++,"-"#" Overlay '%!" ++,"-"." Networks ++,"-"'!" Jo˜ ao Leit˜ ao '$!" Introduction '#!" !"#$%&'()*+$(,%-#./( Overview '!!" CellFarm &!" X-BOT Thicket %!" OpenFire $!" Conclusions #!" Publications !" '" '%!!" ('!!" $%!!" %'!!" )%!!" *'!!" '!%!!" '#'!!" %,+0$1(23(&2%4%$+$%#(52+"-%.(

  62. OpenFire Summary Topology Management for Unstructured Overlay Networks Jo˜ ao Leit˜ ao Introduction OpenFire, based on the enrich technique operating at the Overview peer-to-peer service layer. CellFarm Enrich technique allows to overcome some of the X-BOT challenges introduced by Firewalls and NAT boxes. Thicket OpenFire Conclusions Publications

  63. OpenFire Summary Topology Management for Unstructured Overlay Networks Jo˜ ao Leit˜ ao Introduction OpenFire, based on the enrich technique operating at the Overview peer-to-peer service layer. CellFarm Enrich technique allows to overcome some of the X-BOT challenges introduced by Firewalls and NAT boxes. Thicket OpenFire Conclusions Publications

  64. Roadmap Topology Management for 1 Introduction Unstructured Overlay Networks 2 Overview Jo˜ ao Leit˜ ao Introduction 3 CellFarm Overview CellFarm X-BOT 4 X-BOT Thicket OpenFire 5 Thicket Conclusions Publications 6 OpenFire 7 Conclusions

  65. Conclusions & Future Research Directions Conclusions Topology Management for Unstructured Overlay Networks Jo˜ ao Leit˜ ao 4 techniques to manage the topology of unstructured overlay networks. Introduction Overview Imbuing some relaxed form of structure. CellFarm Protecting the relevant properties of unstructured X-BOT overlays: Thicket Low overhead. OpenFire Robustness to churn scenarios. Conclusions Publications

  66. Conclusions & Future Research Directions Conclusions Topology Management for Unstructured Overlay Networks Jo˜ ao Leit˜ ao 4 techniques to manage the topology of unstructured overlay networks. Introduction Overview Imbuing some relaxed form of structure. CellFarm Protecting the relevant properties of unstructured X-BOT overlays: Thicket Low overhead. OpenFire Robustness to churn scenarios. Conclusions Publications

  67. Conclusions & Future Research Directions Conclusions Topology Management for Unstructured Overlay Networks Jo˜ ao Leit˜ ao At the overlay network layer: Introduction Control (CellFarm). Overview Bias (X-BOT). CellFarm At the peer-to-peer service layer: X-BOT Thicket Embed (Thicket). OpenFire Enrich (OpenFire). Conclusions Publications

  68. Conclusions & Future Research Directions Conclusions Topology Management for Unstructured Overlay Networks In the thesis: Jo˜ ao Leit˜ ao Discusses the design of each solution in detail. Introduction Validates all proposed solutions. Overview Experimentally evaluate their impact over particular case CellFarm studies. X-BOT Summary: The thesis has show and studied how to add Thicket some form structure to unstructured overlay benefiting the OpenFire operation of peer-to-peer services. Conclusions Publications

  69. Conclusions & Future Research Directions Conclusions Topology Management for Unstructured Overlay Networks In the thesis: Jo˜ ao Leit˜ ao Discusses the design of each solution in detail. Introduction Validates all proposed solutions. Overview Experimentally evaluate their impact over particular case CellFarm studies. X-BOT Summary: The thesis has show and studied how to add Thicket some form structure to unstructured overlay benefiting the OpenFire operation of peer-to-peer services. Conclusions Publications

  70. Conclusions & Future Research Directions Future Research Directions Topology Management for Unstructured Overlay Networks Improving volunteer parallel computing platforms Jo˜ ao Leit˜ ao New decentralized volunteer computing infrastructures: Introduction For instance based in solutions similar to CellFarm and Overview X-BOT. CellFarm Offer the opportunity to support more complex parallel X-BOT programing models ( e.g , MapReduce). Thicket Devise a new class of robust and flexible DHT overlay OpenFire Conclusions networks based on the X-BOT design. Publications

  71. Conclusions & Future Research Directions Future Research Directions Topology Management for Unstructured Overlay Networks Improving volunteer parallel computing platforms Jo˜ ao Leit˜ ao New decentralized volunteer computing infrastructures: Introduction For instance based in solutions similar to CellFarm and Overview X-BOT. CellFarm Offer the opportunity to support more complex parallel X-BOT programing models ( e.g , MapReduce). Thicket Devise a new class of robust and flexible DHT overlay OpenFire Conclusions networks based on the X-BOT design. Publications

  72. Conclusions & Future Research Directions Future Research Directions Topology Management for Unstructured Overlay Networks User-centric platform for social applications Jo˜ ao Leit˜ ao Introduction Offer a decentralized platform to support a persistent and Overview dynamic social network. CellFarm Providing the mechanisms for designing new collaborative X-BOT Thicket applications over the social network that can operate in a OpenFire more decentralized fashion. Conclusions Publications

  73. Conclusions & Future Research Directions Future Research Directions Topology Management for Unstructured Overlay Networks User-centric platform for social applications Jo˜ ao Leit˜ ao Introduction Offer a decentralized platform to support a persistent and Overview dynamic social network. CellFarm Providing the mechanisms for designing new collaborative X-BOT Thicket applications over the social network that can operate in a OpenFire more decentralized fashion. Conclusions Publications

  74. Publications Topology Management for Unstructured On the Structure of Unstructured Overlay Networks (fast Overlay Networks abstract) . J. Leit˜ ao, J. Pereira and L. Rodrigues. In Supplement of the 38th Annual IEEE/IFIP International Conference on Dependable Jo˜ ao Leit˜ ao Systems and Networks, Anchorage, Alaska, USA, June, 2008. Introduction Overnesia: a Robust Overlay Network for Virtual Super-Peers . Overview Jo˜ ao Leit˜ ao and Lu´ ıs Rodrigues. Technical Report 36/2009, CellFarm INESC-ID, July 2009 (Available in: X-BOT http: // www. inesc-id. pt/ ficheiros/ publicacoes/ 5510. pdf ). Thicket X-BOT: A Protocol for Resilient Optimization of Unstructured OpenFire Overlays . J. Leit˜ ao, J. P. Marques, J. Pereira and L. Rodrigues. Conclusions Proceedings of the 28th IEEE International Symposium on Reliable Distributed Systems, Niagara Falls, New York, U.S.A., Sep, 2009. pp. Publications 236–245.

  75. Publications Topology Management for Unstructured Overlay On Adding Structure to Unstructured Overlay Networks . J. Leit˜ ao, Networks N. Carvalho, J. Pereira, R. Oliveira, and L. Rodrigues. In Handbook Jo˜ ao Leit˜ ao of Peer-to-Peer Networking, X. Shen, H. Yu, J. Buford, M. Akon (Eds.), Springer 2010. pp. 327-365. ISBN: 978-0-387-09750-3. Introduction Overview Balancing Gossip Exchanges in Networks with Firewalls . J. Leit˜ ao, R. van Renesse and L. Rodrigues. Proceedings of the 9th CellFarm International Workshop on Peer-to-Peer Systems (IPTPS ’10), San X-BOT Jose, CA, USA, 27 April, 2010. Thicket OpenFire Thicket: A Protocol for Building and Maintaining Multiple Trees in a P2P Overlay . M. Ferreira, J. Leit˜ ao, and L. Rodrigues. Conclusions Proceedings of the 29th IEEE Symposium on Reliable Distributed Publications Systems (SRDS), New Delhi, India, 31 October-3 November 2010.

  76. Publications Topology Management for Unstructured Overlay Networks Jo˜ ao Leit˜ ao X-BOT: A Protocol for Resilient Optimization of Unstructured Overlay Networks . J. Leit˜ ao, J. P. Marques, J. Pereira, and L. Introduction Rodrigues. IEEE Transactions on Parallel and Distributed Systems Overview (Published online January 2012). CellFarm Gossip-based Interior-Node-Disjoint Trees . J. Leit˜ ao, M. Ferreira, J. X-BOT Pereira, and L. Rodrigues. IEEE Transactions on Parallel and Thicket Distributed Systems ( Submitted ). OpenFire Conclusions Publications

Recommend

Topology Optimization of an Aircraft Component as a Fluid-Structure System with Unstructured Mesh

Topology Optimization of an Aircraft Component as a Fluid-Structure System with Unstructured Mesh

Topology Optimization of an Aircraft Component as a Fluid-Structure System with Unstructured Mesh Walter J. Paucar Casas 1 , Emanuel M. Cesconeto 2 , Ederval de S. Lisboa, Joo B. D. Moreira, Jos E. Medeiros and Thomas S. Ribeiro 1 Presenter

659 views • 28 slides
ALM API for Topology Management ALM API for Topology Management and Network Layer Transparent

ALM API for Topology Management ALM API for Topology Management and Network Layer Transparent

ALM API for Topology Management ALM API for Topology Management and Network Layer Transparent Multimedia Transport <draft lim irtf sam alm api 00.txt > p 71 ST IETF SAM RG MEETING, PHILADELPHIA. Lim Boon Ping

231 views • 19 slides
Unstructured Data Typically refers to free text I Allows I G Keyword queries including

Unstructured Data Typically refers to free text I Allows I G Keyword queries including

Unstructured Data Management Advanced Topics in Database Management (INFSCI 2711) Textbooks: Database System Concepts - 2010 Introduction to Information Retrieval - 2008 Vladimir Zadorozhny, DINS, SCI, University of Pittsburgh Unstructured Data

349 views • 12 slides
Teaching Unstructured Information Management: Theory and Applications to Computational

Teaching Unstructured Information Management: Theory and Applications to Computational

Teaching Unstructured Information Management: Theory and Applications to Computational Linguistics Students Iryna Gurevych, Christof Mller, Torsten Zesch Ubiquitous Knowledge Processing Group Telecooperation, Computer Science

138 views • 13 slides
UIMA: Unstructured Information Management Architecture Alessandro Moschitti Department of

UIMA: Unstructured Information Management Architecture Alessandro Moschitti Department of

UIMA: Unstructured Information Management Architecture Alessandro Moschitti Department of Computer Science and Information Engineering University of Trento Email: moschitti@disi.unitn.it Motivations Nowadays, natural language processing

620 views • 35 slides
Part III Unstructured Data Data Retrieval: III.1 Unstructured data and data retrieval

Part III Unstructured Data Data Retrieval: III.1 Unstructured data and data retrieval

Inf1-DA 20102011 III: 51 / 89 Part III Unstructured Data Data Retrieval: III.1 Unstructured data and data retrieval Statistical Analysis of Data: III.2 Data scales and summary statistics III.3 Hypothesis testing and correlation III.4

344 views • 17 slides
Part III Unstructured Data Data Retrieval: III.1 Unstructured data and data retrieval

Part III Unstructured Data Data Retrieval: III.1 Unstructured data and data retrieval

Inf1-DA 20102011 III: 28 / 89 Part III Unstructured Data Data Retrieval: III.1 Unstructured data and data retrieval Statistical Analysis of Data: III.2 Data scales and summary statistics III.3 Hypothesis testing and correlation III.4

455 views • 23 slides
Part III Unstructured Data Data Retrieval: III.1 Unstructured data and data retrieval

Part III Unstructured Data Data Retrieval: III.1 Unstructured data and data retrieval

Inf1-DA 20102011 III: 68 / 91 Part III Unstructured Data Data Retrieval: III.1 Unstructured data and data retrieval Statistical Analysis of Data: III.2 Data scales and summary statistics III.3 Hypothesis testing and correlation III.4 2

729 views • 24 slides
Part III Unstructured Data Data Retrieval: III.1 Unstructured data and data retrieval

Part III Unstructured Data Data Retrieval: III.1 Unstructured data and data retrieval

Inf1-DA 20102011 III: 68 / 91 Part III Unstructured Data Data Retrieval: III.1 Unstructured data and data retrieval Statistical Analysis of Data: III.2 Data scales and summary statistics III.3 Hypothesis testing and correlation III.4 2

579 views • 8 slides
WHAT DO WE KNOW ABOUT THE TOPOLOGY? WHAT DO WE KNOW ABOUT THE TOPOLOGY? Number of nodes and

WHAT DO WE KNOW ABOUT THE TOPOLOGY? WHAT DO WE KNOW ABOUT THE TOPOLOGY? Number of nodes and

TxProbe: Discovering Bitcoins Network Topology Using Orphan Transactions Sergi Delgado-Segura , Surya Bakshi, Cristina Prez-Sol, James Litton, Andrew Pachulski, Andrew Miller and Bobby Bhattacharjee sr_gi WHAT DO WE KNOW ABOUT THE TOPOLOGY?

1.84k views • 162 slides
CFD General Notation System (CGNS) Usage for unstructured grids Edwin van der Weide Stanford

CFD General Notation System (CGNS) Usage for unstructured grids Edwin van der Weide Stanford

CFD General Notation System (CGNS) Usage for unstructured grids Edwin van der Weide Stanford University Example Unstructured Grid 2 3 Unstructured grid storage Several possibilities to store an unstructured grid. Every element type

219 views • 19 slides
Enhancements to FreeIPA Replication Topology Management Jan Pazdziora Sr. Principal Software

Enhancements to FreeIPA Replication Topology Management Jan Pazdziora Sr. Principal Software

Enhancements to FreeIPA Replication Topology Management Jan Pazdziora Sr. Principal Software Engineer Identity Management Special Projects, Red Hat 6 th October 2015 FreeIPA Integration of multiple identity-management tools. directory

352 views • 17 slides
Combinatorics and topology of toric arrangements II. Topology of arrangements in the complex torus

Combinatorics and topology of toric arrangements II. Topology of arrangements in the complex torus

Combinatorics and topology of toric arrangements II. Topology of arrangements in the complex torus (An invitation to combinatorial algebraic topology) Emanuele Delucchi (SNSF / Universit e de Fribourg) Toblach/Dobbiaco February 23, 2017

543 views • 42 slides
HUMAN-POWERED DATA MANAGEMENT ! ! Aditya Parameswaran ! ! with H. Garcia-Molina, ! J. Widom, A.

HUMAN-POWERED DATA MANAGEMENT ! ! Aditya Parameswaran ! ! with H. Garcia-Molina, ! J. Widom, A.

HUMAN-POWERED DATA MANAGEMENT ! ! Aditya Parameswaran ! ! with H. Garcia-Molina, ! J. Widom, A. Polyzotis, M. Teh ! ! 1 ! Why should we (DM/DB folks) care? ! Reason 1: Most data is unstructured ! Unstructured Data ! Automated processing: not yet

685 views • 50 slides
73.7% of Healthcare information remains unstructured Healthcare Smart Information Management S

73.7% of Healthcare information remains unstructured Healthcare Smart Information Management S

73.7% of Healthcare information remains unstructured Healthcare Smart Information Management S TRUCTURE P ROCESS A GGREGATE A NALYZE S TRUCTURE P ROCESS A GGREGATE A NALYZE Optimized Reduced Adverse event Data Performance processes wait

381 views • 14 slides
Order Topology Definition Let ( X , < ) be an ordered set. Then the order topology on X is the

Order Topology Definition Let ( X , < ) be an ordered set. Then the order topology on X is the

Order Topology Definition Let ( X , < ) be an ordered set. Then the order topology on X is the topology generated by the basis consisting of unions of sets of the form 1 Open intervals of the form ( a , b ) with a < b in X . 2 If X has a

62 views • 3 slides
Supplementary Information for Article Topology of sustainable management of dynamical systems

Supplementary Information for Article Topology of sustainable management of dynamical systems

Manuscript prepared for Earth Syst. Dynam. with version 2015/09/17 7.94 Copernicus papers of the L A T EX class coperni- cus.cls. Date: 9 November 2015 Supplementary Information for Article Topology of sustainable management of dynamical

54 views • 3 slides
Topological data analysis and topology-based visualization Leila De Floriani Topology-based

Topological data analysis and topology-based visualization Leila De Floriani Topology-based

Topological data analysis and topology-based visualization Leila De Floriani Topology-based methods v Why topology? topology deals with qualitative geometric information , thus v providing a structural description of data topological

298 views • 10 slides
Skill discovery from unstructured demonstrations Skill discovery from unstructured demonstrations

Skill discovery from unstructured demonstrations Skill discovery from unstructured demonstrations

Skill discovery from unstructured demonstrations Skill discovery from unstructured demonstrations Pravesh Ranchod School of Computer Science University of the Witwatersrand pravesh.ranchod@wits.ac.za Initial objective Initial objective We

436 views • 21 slides
Oracle Buys Endeca Combination Creates Best-in-Class Technology and Applications for Unstructured

Oracle Buys Endeca Combination Creates Best-in-Class Technology and Applications for Unstructured

<Insert Picture Here> Oracle Buys Endeca Combination Creates Best-in-Class Technology and Applications for Unstructured Data Management, Business Intelligence and Web Commerce December 5, 2011 Oracle is currently reviewing the existing

325 views • 21 slides
& OTHER QUESTIONS: EQUAL A COFFEE MUG? WHAT IS TOPOLOGY? SARAH BLACKWELL WHAT IS TOPOLOGY?

& OTHER QUESTIONS: EQUAL A COFFEE MUG? WHAT IS TOPOLOGY? SARAH BLACKWELL WHAT IS TOPOLOGY?

WHY DOES A DONUT & OTHER QUESTIONS: EQUAL A COFFEE MUG? WHAT IS TOPOLOGY? SARAH BLACKWELL WHAT IS TOPOLOGY? WHAT DOES IT MEAN FOR TWO THINGS TO BE THE SAME? 1 = 7 (mod 6) color versus colour congruent or similar triangles

542 views • 30 slides
WHAT DO WE KNOW ABOUT THE TOPOLOGY? WHAT DO WE KNOW ABOUT THE TOPOLOGY? Number of nodes and

WHAT DO WE KNOW ABOUT THE TOPOLOGY? WHAT DO WE KNOW ABOUT THE TOPOLOGY? Number of nodes and

# SFBW19 TxProbe: Discovering Bitcoins Network Topology Using Orphan Transactions Sergi Delgado-Segura , Surya Bakshi, Cristina Prez-Sol, James Litton, Andrew Pachulski, Andrew Miller and Bobby Bhattacharjee sr_gi WHAT DO WE KNOW ABOUT THE

1.96k views • 168 slides
Semantic annotation of unstructured and ungrammatical text Matthew Michelson and Craig A.

Semantic annotation of unstructured and ungrammatical text Matthew Michelson and Craig A.

Semantic annotation of unstructured and ungrammatical text Matthew Michelson and Craig A. Knoblock Information Sciences Institute Department of Computer Science University of Southern California Ungrammatical & Unstructured Text

663 views • 29 slides
A Topology-Aware Performance Monitoring Tool for Shared Resource Management in Multicore Systems

A Topology-Aware Performance Monitoring Tool for Shared Resource Management in Multicore Systems

A Topology-Aware Performance Monitoring Tool for Shared Resource Management in Multicore Systems TADaaM Team - Nicolas Denoyelle - Brice Goglin - Emmanuel Jeannot August 24, 2015 1. Context/Motivations 2. Fast presentation of the tool 3.

446 views • 29 slides

More recommend