SlideShare uma empresa Scribd logo
1 de 80
Baixar para ler offline
Motivation
                       Algorithms
                  Our Contribution
                       Evaluation
                       Conclusion




Scaling Online Social Networks: extended SPAR
             using Gossip Learning

     Presented by: Muhammad Anis uddin Nasir
              Coworker: Maria Stylianou
         Supervised by: Sarunas Girdzijauskas

                KTH Royal Institute of Technology


                      December 5, 2012



        Muhammad Anis uddin Nasir    Scaling Online Social Networks   1/24
Motivation
                                Algorithms
                           Our Contribution
                                Evaluation
                                Conclusion



1   Motivation
2   Algorithms
      SPAR
      JA-BE-JA
3   Our Contribution
      Challenges
      Proposed Algorithm
4   Evaluation
      Datasets
      Implementation
      Results
5   Conclusion

                 Muhammad Anis uddin Nasir    Scaling Online Social Networks   2/24
Motivation
                           Algorithms
                      Our Contribution
                           Evaluation
                           Conclusion


Online Social Networks




            Muhammad Anis uddin Nasir    Scaling Online Social Networks   3/24
Motivation
                              Algorithms
                         Our Contribution
                              Evaluation
                              Conclusion


Scalability




     Hardware Scalability




               Muhammad Anis uddin Nasir    Scaling Online Social Networks   4/24
Motivation
                              Algorithms
                         Our Contribution
                              Evaluation
                              Conclusion


Scalability




     Hardware Scalability
     Application Scalability




               Muhammad Anis uddin Nasir    Scaling Online Social Networks   4/24
Motivation
                             Algorithms
                        Our Contribution
                             Evaluation
                             Conclusion


Scaling



    Vertical Scaling




              Muhammad Anis uddin Nasir    Scaling Online Social Networks   5/24
Motivation
                                Algorithms
                           Our Contribution
                                Evaluation
                                Conclusion


Scaling



    Vertical Scaling
          Full Replication




                 Muhammad Anis uddin Nasir    Scaling Online Social Networks   5/24
Motivation
                                Algorithms
                           Our Contribution
                                Evaluation
                                Conclusion


Scaling



    Vertical Scaling
          Full Replication
          Data Locality




                 Muhammad Anis uddin Nasir    Scaling Online Social Networks   5/24
Motivation
                                Algorithms
                           Our Contribution
                                Evaluation
                                Conclusion


Scaling



    Vertical Scaling
          Full Replication
          Data Locality
          High Cost




                 Muhammad Anis uddin Nasir    Scaling Online Social Networks   5/24
Motivation
                                Algorithms
                           Our Contribution
                                Evaluation
                                Conclusion


Scaling



    Vertical Scaling
          Full Replication
          Data Locality
          High Cost
    Horizontal Scaling




                 Muhammad Anis uddin Nasir    Scaling Online Social Networks   5/24
Motivation
                                Algorithms
                           Our Contribution
                                Evaluation
                                Conclusion


Scaling



    Vertical Scaling
          Full Replication
          Data Locality
          High Cost
    Horizontal Scaling
          Sharding




                 Muhammad Anis uddin Nasir    Scaling Online Social Networks   5/24
Motivation
                                Algorithms
                           Our Contribution
                                Evaluation
                                Conclusion


Scaling



    Vertical Scaling
          Full Replication
          Data Locality
          High Cost
    Horizontal Scaling
          Sharding
          Disjoint Data




                 Muhammad Anis uddin Nasir    Scaling Online Social Networks   5/24
Motivation
                                Algorithms
                           Our Contribution
                                Evaluation
                                Conclusion


Scaling



    Vertical Scaling
          Full Replication
          Data Locality
          High Cost
    Horizontal Scaling
          Sharding
          Disjoint Data
          Partitioning OSNs




                 Muhammad Anis uddin Nasir    Scaling Online Social Networks   5/24
Motivation
                                Algorithms
                                              SPAR
                           Our Contribution
                                              JA-BE-JA
                                Evaluation
                                Conclusion



1   Motivation
2   Algorithms
      SPAR
      JA-BE-JA
3   Our Contribution
      Challenges
      Proposed Algorithm
4   Evaluation
      Datasets
      Implementation
      Results
5   Conclusion

                 Muhammad Anis uddin Nasir    Scaling Online Social Networks   6/24
Motivation
                            Algorithms
                                          SPAR
                       Our Contribution
                                          JA-BE-JA
                            Evaluation
                            Conclusion


Features



    Local Semantics




             Muhammad Anis uddin Nasir    Scaling Online Social Networks   7/24
Motivation
                            Algorithms
                                          SPAR
                       Our Contribution
                                          JA-BE-JA
                            Evaluation
                            Conclusion


Features



    Local Semantics

    Load Balancing




             Muhammad Anis uddin Nasir    Scaling Online Social Networks   7/24
Motivation
                             Algorithms
                                           SPAR
                        Our Contribution
                                           JA-BE-JA
                             Evaluation
                             Conclusion


Features



    Local Semantics

    Load Balancing

    Fault Tolerant




              Muhammad Anis uddin Nasir    Scaling Online Social Networks   7/24
Motivation
                             Algorithms
                                           SPAR
                        Our Contribution
                                           JA-BE-JA
                             Evaluation
                             Conclusion


Features



    Local Semantics

    Load Balancing

    Fault Tolerant

    Dynamic




              Muhammad Anis uddin Nasir    Scaling Online Social Networks   7/24
Motivation
                             Algorithms
                                           SPAR
                        Our Contribution
                                           JA-BE-JA
                             Evaluation
                             Conclusion


Features



    Local Semantics

    Load Balancing

    Fault Tolerant

    Dynamic

    Low Replication Overhead




              Muhammad Anis uddin Nasir    Scaling Online Social Networks   7/24
Motivation
                              Algorithms
                                            SPAR
                         Our Contribution
                                            JA-BE-JA
                              Evaluation
                              Conclusion


Architecture


    Partition Manager




               Muhammad Anis uddin Nasir    Scaling Online Social Networks   8/24
Motivation
                              Algorithms
                                            SPAR
                         Our Contribution
                                            JA-BE-JA
                              Evaluation
                              Conclusion


Architecture


    Partition Manager


    Directory Service




               Muhammad Anis uddin Nasir    Scaling Online Social Networks   8/24
Motivation
                              Algorithms
                                            SPAR
                         Our Contribution
                                            JA-BE-JA
                              Evaluation
                              Conclusion


Architecture


    Partition Manager


    Directory Service


    Local Directory
    Service




               Muhammad Anis uddin Nasir    Scaling Online Social Networks   8/24
Motivation
                              Algorithms
                                            SPAR
                         Our Contribution
                                            JA-BE-JA
                              Evaluation
                              Conclusion


Architecture


    Partition Manager


    Directory Service


    Local Directory
    Service


    Replication
    Manager


               Muhammad Anis uddin Nasir    Scaling Online Social Networks   8/24
Motivation
                               Algorithms
                                             SPAR
                          Our Contribution
                                             JA-BE-JA
                               Evaluation
                               Conclusion


SPAR Algorithm


    Heuristic




                Muhammad Anis uddin Nasir    Scaling Online Social Networks   9/24
Motivation
                               Algorithms
                                             SPAR
                          Our Contribution
                                             JA-BE-JA
                               Evaluation
                               Conclusion


SPAR Algorithm


    Heuristic
        Greedy Optimization




                Muhammad Anis uddin Nasir    Scaling Online Social Networks   9/24
Motivation
                               Algorithms
                                             SPAR
                          Our Contribution
                                             JA-BE-JA
                               Evaluation
                               Conclusion


SPAR Algorithm


    Heuristic
        Greedy Optimization
        Local Search




                Muhammad Anis uddin Nasir    Scaling Online Social Networks   9/24
Motivation
                               Algorithms
                                             SPAR
                          Our Contribution
                                             JA-BE-JA
                               Evaluation
                               Conclusion


SPAR Algorithm


    Heuristic
        Greedy Optimization
        Local Search

    Node Add/Remove




                Muhammad Anis uddin Nasir    Scaling Online Social Networks   9/24
Motivation
                               Algorithms
                                             SPAR
                          Our Contribution
                                             JA-BE-JA
                               Evaluation
                               Conclusion


SPAR Algorithm


    Heuristic
        Greedy Optimization
        Local Search

    Node Add/Remove

    Edge Add/Remove




                Muhammad Anis uddin Nasir    Scaling Online Social Networks   9/24
Motivation
                               Algorithms
                                             SPAR
                          Our Contribution
                                             JA-BE-JA
                               Evaluation
                               Conclusion


SPAR Algorithm


    Heuristic
        Greedy Optimization
        Local Search

    Node Add/Remove

    Edge Add/Remove
        3 Configurations




                Muhammad Anis uddin Nasir    Scaling Online Social Networks   9/24
Motivation
                               Algorithms
                                             SPAR
                          Our Contribution
                                             JA-BE-JA
                               Evaluation
                               Conclusion


SPAR Algorithm


    Heuristic
        Greedy Optimization
        Local Search

    Node Add/Remove

    Edge Add/Remove
        3 Configurations

    Server Add/Remove




                Muhammad Anis uddin Nasir    Scaling Online Social Networks   9/24
Motivation
                               Algorithms
                                             SPAR
                          Our Contribution
                                             JA-BE-JA
                               Evaluation
                               Conclusion


SPAR Algorithm


    Heuristic
        Greedy Optimization
        Local Search

    Node Add/Remove

    Edge Add/Remove
        3 Configurations

    Server Add/Remove
        Redistribution



                Muhammad Anis uddin Nasir    Scaling Online Social Networks   9/24
Motivation
                               Algorithms
                                             SPAR
                          Our Contribution
                                             JA-BE-JA
                               Evaluation
                               Conclusion


SPAR Algorithm


    Heuristic
        Greedy Optimization
        Local Search

    Node Add/Remove

    Edge Add/Remove
        3 Configurations

    Server Add/Remove
        Redistribution
        Let it fill


                Muhammad Anis uddin Nasir    Scaling Online Social Networks   9/24
Motivation
                               Algorithms
                                             SPAR
                          Our Contribution
                                             JA-BE-JA
                               Evaluation
                               Conclusion


SPAR Algorithm


    Heuristic
        Greedy Optimization
        Local Search

    Node Add/Remove

    Edge Add/Remove
        3 Configurations

    Server Add/Remove
        Redistribution
        Let it fill


                Muhammad Anis uddin Nasir    Scaling Online Social Networks   9/24
Motivation
                               Algorithms
                                             SPAR
                          Our Contribution
                                             JA-BE-JA
                               Evaluation
                               Conclusion


SPAR Algorithm


    Heuristic
        Greedy Optimization
        Local Search

    Node Add/Remove

    Edge Add/Remove
        3 Configurations

    Server Add/Remove
        Redistribution
        Let it fill


                Muhammad Anis uddin Nasir    Scaling Online Social Networks   9/24
Motivation
                               Algorithms
                                             SPAR
                          Our Contribution
                                             JA-BE-JA
                               Evaluation
                               Conclusion


SPAR Algorithm


    Heuristic
        Greedy Optimization
        Local Search

    Node Add/Remove

    Edge Add/Remove
        3 Configurations

    Server Add/Remove
        Redistribution
        Let it fill


                Muhammad Anis uddin Nasir    Scaling Online Social Networks   9/24
Motivation
                               Algorithms
                                             SPAR
                          Our Contribution
                                             JA-BE-JA
                               Evaluation
                               Conclusion


SPAR Algorithm


    Heuristic
        Greedy Optimization
        Local Search

    Node Add/Remove

    Edge Add/Remove
        3 Configurations

    Server Add/Remove
        Redistribution
        Let it fill


                Muhammad Anis uddin Nasir    Scaling Online Social Networks   9/24
Motivation
                             Algorithms
                                           SPAR
                        Our Contribution
                                           JA-BE-JA
                             Evaluation
                             Conclusion


Features


    Distributed Partitioning




              Muhammad Anis uddin Nasir    Scaling Online Social Networks   10/24
Motivation
                             Algorithms
                                           SPAR
                        Our Contribution
                                           JA-BE-JA
                             Evaluation
                             Conclusion


Features


    Distributed Partitioning

    k-way Partitioning




              Muhammad Anis uddin Nasir    Scaling Online Social Networks   10/24
Motivation
                             Algorithms
                                           SPAR
                        Our Contribution
                                           JA-BE-JA
                             Evaluation
                             Conclusion


Features


    Distributed Partitioning

    k-way Partitioning

    Load Balancing




              Muhammad Anis uddin Nasir    Scaling Online Social Networks   10/24
Motivation
                             Algorithms
                                           SPAR
                        Our Contribution
                                           JA-BE-JA
                             Evaluation
                             Conclusion


Features


    Distributed Partitioning

    k-way Partitioning

    Load Balancing

    Low Inter-communication
    Overhead




              Muhammad Anis uddin Nasir    Scaling Online Social Networks   10/24
Motivation
                             Algorithms
                                           SPAR
                        Our Contribution
                                           JA-BE-JA
                             Evaluation
                             Conclusion


Features


    Distributed Partitioning

    k-way Partitioning

    Load Balancing

    Low Inter-communication
    Overhead

    Local Search



              Muhammad Anis uddin Nasir    Scaling Online Social Networks   10/24
Motivation
                             Algorithms
                                           SPAR
                        Our Contribution
                                           JA-BE-JA
                             Evaluation
                             Conclusion


Overview

    Sampling Policies




              Muhammad Anis uddin Nasir    Scaling Online Social Networks   11/24
Motivation
                               Algorithms
                                             SPAR
                          Our Contribution
                                             JA-BE-JA
                               Evaluation
                               Conclusion


Overview

    Sampling Policies
        Local




                Muhammad Anis uddin Nasir    Scaling Online Social Networks   11/24
Motivation
                             Algorithms
                                           SPAR
                        Our Contribution
                                           JA-BE-JA
                             Evaluation
                             Conclusion


Overview

    Sampling Policies
        Local
        Random




              Muhammad Anis uddin Nasir    Scaling Online Social Networks   11/24
Motivation
                             Algorithms
                                           SPAR
                        Our Contribution
                                           JA-BE-JA
                             Evaluation
                             Conclusion


Overview

    Sampling Policies
        Local
        Random
        Hybrid




              Muhammad Anis uddin Nasir    Scaling Online Social Networks   11/24
Motivation
                             Algorithms
                                           SPAR
                        Our Contribution
                                           JA-BE-JA
                             Evaluation
                             Conclusion


Overview

    Sampling Policies
        Local
        Random
        Hybrid

    Swapping Policies




              Muhammad Anis uddin Nasir    Scaling Online Social Networks   11/24
Motivation
                             Algorithms
                                           SPAR
                        Our Contribution
                                           JA-BE-JA
                             Evaluation
                             Conclusion


Overview

    Sampling Policies
        Local
        Random
        Hybrid

    Swapping Policies
        Energy Function




              Muhammad Anis uddin Nasir    Scaling Online Social Networks   11/24
Motivation
                             Algorithms
                                           SPAR
                        Our Contribution
                                           JA-BE-JA
                             Evaluation
                             Conclusion


Overview

    Sampling Policies
        Local
        Random
        Hybrid

    Swapping Policies
        Energy Function
        Simulated Annealing




              Muhammad Anis uddin Nasir    Scaling Online Social Networks   11/24
Motivation
                               Algorithms
                                             SPAR
                          Our Contribution
                                             JA-BE-JA
                               Evaluation
                               Conclusion


Overview

    Sampling Policies
        Local
        Random
        Hybrid

    Swapping Policies
        Energy Function
        Simulated Annealing

    Algorithm




                Muhammad Anis uddin Nasir    Scaling Online Social Networks   11/24
Motivation
                               Algorithms
                                             SPAR
                          Our Contribution
                                             JA-BE-JA
                               Evaluation
                               Conclusion


Overview

    Sampling Policies
        Local
        Random
        Hybrid

    Swapping Policies
        Energy Function
        Simulated Annealing

    Algorithm
        Hybrid Sampling



                Muhammad Anis uddin Nasir    Scaling Online Social Networks   11/24
Motivation
                               Algorithms
                                             SPAR
                          Our Contribution
                                             JA-BE-JA
                               Evaluation
                               Conclusion


Overview

    Sampling Policies
        Local
        Random
        Hybrid

    Swapping Policies
        Energy Function
        Simulated Annealing

    Algorithm
        Hybrid Sampling
        Simulated Annealing


                Muhammad Anis uddin Nasir    Scaling Online Social Networks   11/24
Motivation
                                Algorithms
                                              Challenges
                           Our Contribution
                                              Proposed Algorithm
                                Evaluation
                                Conclusion



1   Motivation
2   Algorithms
      SPAR
      JA-BE-JA
3   Our Contribution
      Challenges
      Proposed Algorithm
4   Evaluation
      Datasets
      Implementation
      Results
5   Conclusion

                 Muhammad Anis uddin Nasir    Scaling Online Social Networks   12/24
Motivation
                            Algorithms
                                          Challenges
                       Our Contribution
                                          Proposed Algorithm
                            Evaluation
                            Conclusion


Challenges


    Global View




             Muhammad Anis uddin Nasir    Scaling Online Social Networks   13/24
Motivation
                            Algorithms
                                          Challenges
                       Our Contribution
                                          Proposed Algorithm
                            Evaluation
                            Conclusion


Challenges


    Global View


    Partition Manager




             Muhammad Anis uddin Nasir    Scaling Online Social Networks   13/24
Motivation
                            Algorithms
                                          Challenges
                       Our Contribution
                                          Proposed Algorithm
                            Evaluation
                            Conclusion


Challenges


    Global View


    Partition Manager


    Replication Overhead




             Muhammad Anis uddin Nasir    Scaling Online Social Networks   13/24
Motivation
                            Algorithms
                                          Challenges
                       Our Contribution
                                          Proposed Algorithm
                            Evaluation
                            Conclusion


Challenges


    Global View


    Partition Manager


    Replication Overhead


    Load Balancing



             Muhammad Anis uddin Nasir    Scaling Online Social Networks   13/24
Motivation
                           Algorithms
                                         Challenges
                      Our Contribution
                                         Proposed Algorithm
                           Evaluation
                           Conclusion


Proposed Algorithm



    SPAR + JA-BE-JA




            Muhammad Anis uddin Nasir    Scaling Online Social Networks   14/24
Motivation
                            Algorithms
                                          Challenges
                       Our Contribution
                                          Proposed Algorithm
                            Evaluation
                            Conclusion


Proposed Algorithm



    SPAR + JA-BE-JA

    Gossip Learning




             Muhammad Anis uddin Nasir    Scaling Online Social Networks   14/24
Motivation
                               Algorithms
                                             Challenges
                          Our Contribution
                                             Proposed Algorithm
                               Evaluation
                               Conclusion


Proposed Algorithm



    SPAR + JA-BE-JA

    Gossip Learning

    Simulated
    Annealing




                Muhammad Anis uddin Nasir    Scaling Online Social Networks   14/24
Motivation
                               Algorithms
                                             Challenges
                          Our Contribution
                                             Proposed Algorithm
                               Evaluation
                               Conclusion


Proposed Algorithm



    SPAR + JA-BE-JA

    Gossip Learning

    Simulated
    Annealing

    Optimal Replication




                Muhammad Anis uddin Nasir    Scaling Online Social Networks   14/24
Motivation
                                Algorithms    Datasets
                           Our Contribution   Implementation
                                Evaluation    Results
                                Conclusion



1   Motivation
2   Algorithms
      SPAR
      JA-BE-JA
3   Our Contribution
      Challenges
      Proposed Algorithm
4   Evaluation
      Datasets
      Implementation
      Results
5   Conclusion

                 Muhammad Anis uddin Nasir    Scaling Online Social Networks   15/24
Motivation
                                  Algorithms    Datasets
                             Our Contribution   Implementation
                                  Evaluation    Results
                                  Conclusion


Datasets




    Synthetic Graphs




    0
        http://snap.stanford.edu/data/
                   Muhammad Anis uddin Nasir    Scaling Online Social Networks   16/24
Motivation
                                  Algorithms    Datasets
                             Our Contribution   Implementation
                                  Evaluation    Results
                                  Conclusion


Datasets




    Synthetic Graphs


    Facebook Graphs




    0
        http://snap.stanford.edu/data/
                   Muhammad Anis uddin Nasir    Scaling Online Social Networks   16/24
Motivation
                                        Algorithms    Datasets
                                   Our Contribution   Implementation
                                        Evaluation    Results
                                        Conclusion


Datasets


    Synthetic Graphs
              Randomized




    Facebook Graphs




    0
        https://gephi.org/
                         Muhammad Anis uddin Nasir    Scaling Online Social Networks   17/24
Motivation
                                        Algorithms    Datasets
                                   Our Contribution   Implementation
                                        Evaluation    Results
                                        Conclusion


Datasets


    Synthetic Graphs
              Randomized
              Clustered



    Facebook Graphs




    0
        https://gephi.org/
                         Muhammad Anis uddin Nasir    Scaling Online Social Networks   17/24
Motivation
                                        Algorithms    Datasets
                                   Our Contribution   Implementation
                                        Evaluation    Results
                                        Conclusion


Datasets


    Synthetic Graphs
              Randomized
              Clustered
              Highly Clustered


    Facebook Graphs




    0
        https://gephi.org/
                         Muhammad Anis uddin Nasir    Scaling Online Social Networks   17/24
Motivation
                                        Algorithms    Datasets
                                   Our Contribution   Implementation
                                        Evaluation    Results
                                        Conclusion


Datasets


    Synthetic Graphs
              Randomized
              Clustered
              Highly Clustered


    Facebook Graphs
              150 nodes, 3386 edges




    0
        https://gephi.org/
                         Muhammad Anis uddin Nasir    Scaling Online Social Networks   17/24
Motivation
                                        Algorithms    Datasets
                                   Our Contribution   Implementation
                                        Evaluation    Results
                                        Conclusion


Datasets


    Synthetic Graphs
              Randomized
              Clustered
              Highly Clustered


    Facebook Graphs
              150 nodes, 3386 edges
              224 nodes, 6384 edges




    0
        https://gephi.org/
                         Muhammad Anis uddin Nasir    Scaling Online Social Networks   17/24
Motivation
                                        Algorithms    Datasets
                                   Our Contribution   Implementation
                                        Evaluation    Results
                                        Conclusion


Datasets


    Synthetic Graphs
              Randomized
              Clustered
              Highly Clustered


    Facebook Graphs
              150 nodes, 3386 edges
              224 nodes, 6384 edges
              786 nodes, 60050 edges


    0
        https://gephi.org/
                         Muhammad Anis uddin Nasir    Scaling Online Social Networks   17/24
Motivation
                            Algorithms    Datasets
                       Our Contribution   Implementation
                            Evaluation    Results
                            Conclusion


Implementation



    SPAR


    Proposed Algorithm


    Metric
       Replication Overhead




             Muhammad Anis uddin Nasir    Scaling Online Social Networks   18/24
Motivation
                                          Algorithms        Datasets
                                     Our Contribution       Implementation
                                          Evaluation        Results
                                          Conclusion


Evaluation of Replication Overhead
         Replication Factor
                 Fault tolerance replicas reduce replication overhead
                 Proposed Algorithm performs better than SPAR




    0
        replication overhead = number of replicas/number of users
                         Muhammad Anis uddin Nasir          Scaling Online Social Networks   19/24
Motivation
                                          Algorithms        Datasets
                                     Our Contribution       Implementation
                                          Evaluation        Results
                                          Conclusion


Evaluation of Replication Overhead

         Number of Servers
                 Less Replication overhead in the case of proposed algorithm
                 Proposed Algorithm performs better in the case of high
                 clusterization




    0
        replication overhead = number of replicas/number of users
                         Muhammad Anis uddin Nasir          Scaling Online Social Networks   20/24
Motivation
                                          Algorithms        Datasets
                                     Our Contribution       Implementation
                                          Evaluation        Results
                                          Conclusion


Evaluation of Replication Overhead

         Number of Servers
                 Less Replication overhead in the case of proposed algorithm
                 Proposed Algorithm performs better in case of high
                 clusterization




    0
        replication overhead = number of replicas/number of users
                         Muhammad Anis uddin Nasir          Scaling Online Social Networks   21/24
Motivation
                                Algorithms
                           Our Contribution
                                Evaluation
                                Conclusion



1   Motivation
2   Algorithms
      SPAR
      JA-BE-JA
3   Our Contribution
      Challenges
      Proposed Algorithm
4   Evaluation
      Datasets
      Implementation
      Results
5   Conclusion

                 Muhammad Anis uddin Nasir    Scaling Online Social Networks   22/24
Motivation
                            Algorithms
                       Our Contribution
                            Evaluation
                            Conclusion


Conclusion


     Distributed social-based partitioning




             Muhammad Anis uddin Nasir    Scaling Online Social Networks   23/24
Motivation
                            Algorithms
                       Our Contribution
                            Evaluation
                            Conclusion


Conclusion


     Distributed social-based partitioning

     Local Semantics




             Muhammad Anis uddin Nasir    Scaling Online Social Networks   23/24
Motivation
                            Algorithms
                       Our Contribution
                            Evaluation
                            Conclusion


Conclusion


     Distributed social-based partitioning

     Local Semantics

     Reduced Replication overhead compared to SPAR




             Muhammad Anis uddin Nasir    Scaling Online Social Networks   23/24
Motivation
                            Algorithms
                       Our Contribution
                            Evaluation
                            Conclusion


Conclusion


     Distributed social-based partitioning

     Local Semantics

     Reduced Replication overhead compared to SPAR

     Better load balancing using k-way partitioning




             Muhammad Anis uddin Nasir    Scaling Online Social Networks   23/24
Motivation
                            Algorithms
                       Our Contribution
                            Evaluation
                            Conclusion


Conclusion


     Distributed social-based partitioning

     Local Semantics

     Reduced Replication overhead compared to SPAR

     Better load balancing using k-way partitioning

     Transparent Scaling




             Muhammad Anis uddin Nasir    Scaling Online Social Networks   23/24
Motivation
                       Algorithms
                  Our Contribution
                       Evaluation
                       Conclusion




Scaling Online Social Networks: extended SPAR
             using Gossip Learning

     Presented by: Muhammad Anis uddin Nasir
              Coworker: Maria Stylianou
         Supervised by: Sarunas Girdzijauskas

                KTH Royal Institute of Technology


                      December 5, 2012



        Muhammad Anis uddin Nasir    Scaling Online Social Networks   24/24

Mais conteúdo relacionado

Mais de Anis Nasir

The Power of Both Choices: Practical Load Balancing for Distributed Stream Pr...
The Power of Both Choices: Practical Load Balancing for Distributed Stream Pr...The Power of Both Choices: Practical Load Balancing for Distributed Stream Pr...
The Power of Both Choices: Practical Load Balancing for Distributed Stream Pr...Anis Nasir
 
Gossip based partitioning and replication for Online Social Networks
Gossip based partitioning and replication for Online Social NetworksGossip based partitioning and replication for Online Social Networks
Gossip based partitioning and replication for Online Social NetworksAnis Nasir
 
Pushing the cap
Pushing the capPushing the cap
Pushing the capAnis Nasir
 
Final presentation survey p2p-videostreamingwmn
Final presentation survey p2p-videostreamingwmnFinal presentation survey p2p-videostreamingwmn
Final presentation survey p2p-videostreamingwmnAnis Nasir
 

Mais de Anis Nasir (8)

The Power of Both Choices: Practical Load Balancing for Distributed Stream Pr...
The Power of Both Choices: Practical Load Balancing for Distributed Stream Pr...The Power of Both Choices: Practical Load Balancing for Distributed Stream Pr...
The Power of Both Choices: Practical Load Balancing for Distributed Stream Pr...
 
Gossip based partitioning and replication for Online Social Networks
Gossip based partitioning and replication for Online Social NetworksGossip based partitioning and replication for Online Social Networks
Gossip based partitioning and replication for Online Social Networks
 
Pushing the cap
Pushing the capPushing the cap
Pushing the cap
 
Mesos
MesosMesos
Mesos
 
NaaS
NaaSNaaS
NaaS
 
NaaS
NaaSNaaS
NaaS
 
Final presentation survey p2p-videostreamingwmn
Final presentation survey p2p-videostreamingwmnFinal presentation survey p2p-videostreamingwmn
Final presentation survey p2p-videostreamingwmn
 
RESTvsSOAP
RESTvsSOAPRESTvsSOAP
RESTvsSOAP
 

Último

Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...Drew Madelung
 
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerHow to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerThousandEyes
 
08448380779 Call Girls In Friends Colony Women Seeking Men
08448380779 Call Girls In Friends Colony Women Seeking Men08448380779 Call Girls In Friends Colony Women Seeking Men
08448380779 Call Girls In Friends Colony Women Seeking MenDelhi Call girls
 
My Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationMy Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationRidwan Fadjar
 
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...Miguel Araújo
 
SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024Scott Keck-Warren
 
Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Allon Mureinik
 
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptxHampshireHUG
 
Unblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen FramesUnblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen FramesSinan KOZAK
 
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdfThe Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdfEnterprise Knowledge
 
Salesforce Community Group Quito, Salesforce 101
Salesforce Community Group Quito, Salesforce 101Salesforce Community Group Quito, Salesforce 101
Salesforce Community Group Quito, Salesforce 101Paola De la Torre
 
CNv6 Instructor Chapter 6 Quality of Service
CNv6 Instructor Chapter 6 Quality of ServiceCNv6 Instructor Chapter 6 Quality of Service
CNv6 Instructor Chapter 6 Quality of Servicegiselly40
 
08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking MenDelhi Call girls
 
🐬 The future of MySQL is Postgres 🐘
🐬  The future of MySQL is Postgres   🐘🐬  The future of MySQL is Postgres   🐘
🐬 The future of MySQL is Postgres 🐘RTylerCroy
 
Slack Application Development 101 Slides
Slack Application Development 101 SlidesSlack Application Development 101 Slides
Slack Application Development 101 Slidespraypatel2
 
Google AI Hackathon: LLM based Evaluator for RAG
Google AI Hackathon: LLM based Evaluator for RAGGoogle AI Hackathon: LLM based Evaluator for RAG
Google AI Hackathon: LLM based Evaluator for RAGSujit Pal
 
The Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxThe Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxMalak Abu Hammad
 
Swan(sea) Song – personal research during my six years at Swansea ... and bey...
Swan(sea) Song – personal research during my six years at Swansea ... and bey...Swan(sea) Song – personal research during my six years at Swansea ... and bey...
Swan(sea) Song – personal research during my six years at Swansea ... and bey...Alan Dix
 
Understanding the Laravel MVC Architecture
Understanding the Laravel MVC ArchitectureUnderstanding the Laravel MVC Architecture
Understanding the Laravel MVC ArchitecturePixlogix Infotech
 
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024BookNet Canada
 

Último (20)

Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
 
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerHow to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected Worker
 
08448380779 Call Girls In Friends Colony Women Seeking Men
08448380779 Call Girls In Friends Colony Women Seeking Men08448380779 Call Girls In Friends Colony Women Seeking Men
08448380779 Call Girls In Friends Colony Women Seeking Men
 
My Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationMy Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 Presentation
 
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
 
SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024
 
Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)
 
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
 
Unblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen FramesUnblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen Frames
 
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdfThe Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
 
Salesforce Community Group Quito, Salesforce 101
Salesforce Community Group Quito, Salesforce 101Salesforce Community Group Quito, Salesforce 101
Salesforce Community Group Quito, Salesforce 101
 
CNv6 Instructor Chapter 6 Quality of Service
CNv6 Instructor Chapter 6 Quality of ServiceCNv6 Instructor Chapter 6 Quality of Service
CNv6 Instructor Chapter 6 Quality of Service
 
08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men
 
🐬 The future of MySQL is Postgres 🐘
🐬  The future of MySQL is Postgres   🐘🐬  The future of MySQL is Postgres   🐘
🐬 The future of MySQL is Postgres 🐘
 
Slack Application Development 101 Slides
Slack Application Development 101 SlidesSlack Application Development 101 Slides
Slack Application Development 101 Slides
 
Google AI Hackathon: LLM based Evaluator for RAG
Google AI Hackathon: LLM based Evaluator for RAGGoogle AI Hackathon: LLM based Evaluator for RAG
Google AI Hackathon: LLM based Evaluator for RAG
 
The Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxThe Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptx
 
Swan(sea) Song – personal research during my six years at Swansea ... and bey...
Swan(sea) Song – personal research during my six years at Swansea ... and bey...Swan(sea) Song – personal research during my six years at Swansea ... and bey...
Swan(sea) Song – personal research during my six years at Swansea ... and bey...
 
Understanding the Laravel MVC Architecture
Understanding the Laravel MVC ArchitectureUnderstanding the Laravel MVC Architecture
Understanding the Laravel MVC Architecture
 
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
 

Scaling Online Social Networks: extended SPAR using Gossip Learning

  • 1. Motivation Algorithms Our Contribution Evaluation Conclusion Scaling Online Social Networks: extended SPAR using Gossip Learning Presented by: Muhammad Anis uddin Nasir Coworker: Maria Stylianou Supervised by: Sarunas Girdzijauskas KTH Royal Institute of Technology December 5, 2012 Muhammad Anis uddin Nasir Scaling Online Social Networks 1/24
  • 2. Motivation Algorithms Our Contribution Evaluation Conclusion 1 Motivation 2 Algorithms SPAR JA-BE-JA 3 Our Contribution Challenges Proposed Algorithm 4 Evaluation Datasets Implementation Results 5 Conclusion Muhammad Anis uddin Nasir Scaling Online Social Networks 2/24
  • 3. Motivation Algorithms Our Contribution Evaluation Conclusion Online Social Networks Muhammad Anis uddin Nasir Scaling Online Social Networks 3/24
  • 4. Motivation Algorithms Our Contribution Evaluation Conclusion Scalability Hardware Scalability Muhammad Anis uddin Nasir Scaling Online Social Networks 4/24
  • 5. Motivation Algorithms Our Contribution Evaluation Conclusion Scalability Hardware Scalability Application Scalability Muhammad Anis uddin Nasir Scaling Online Social Networks 4/24
  • 6. Motivation Algorithms Our Contribution Evaluation Conclusion Scaling Vertical Scaling Muhammad Anis uddin Nasir Scaling Online Social Networks 5/24
  • 7. Motivation Algorithms Our Contribution Evaluation Conclusion Scaling Vertical Scaling Full Replication Muhammad Anis uddin Nasir Scaling Online Social Networks 5/24
  • 8. Motivation Algorithms Our Contribution Evaluation Conclusion Scaling Vertical Scaling Full Replication Data Locality Muhammad Anis uddin Nasir Scaling Online Social Networks 5/24
  • 9. Motivation Algorithms Our Contribution Evaluation Conclusion Scaling Vertical Scaling Full Replication Data Locality High Cost Muhammad Anis uddin Nasir Scaling Online Social Networks 5/24
  • 10. Motivation Algorithms Our Contribution Evaluation Conclusion Scaling Vertical Scaling Full Replication Data Locality High Cost Horizontal Scaling Muhammad Anis uddin Nasir Scaling Online Social Networks 5/24
  • 11. Motivation Algorithms Our Contribution Evaluation Conclusion Scaling Vertical Scaling Full Replication Data Locality High Cost Horizontal Scaling Sharding Muhammad Anis uddin Nasir Scaling Online Social Networks 5/24
  • 12. Motivation Algorithms Our Contribution Evaluation Conclusion Scaling Vertical Scaling Full Replication Data Locality High Cost Horizontal Scaling Sharding Disjoint Data Muhammad Anis uddin Nasir Scaling Online Social Networks 5/24
  • 13. Motivation Algorithms Our Contribution Evaluation Conclusion Scaling Vertical Scaling Full Replication Data Locality High Cost Horizontal Scaling Sharding Disjoint Data Partitioning OSNs Muhammad Anis uddin Nasir Scaling Online Social Networks 5/24
  • 14. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion 1 Motivation 2 Algorithms SPAR JA-BE-JA 3 Our Contribution Challenges Proposed Algorithm 4 Evaluation Datasets Implementation Results 5 Conclusion Muhammad Anis uddin Nasir Scaling Online Social Networks 6/24
  • 15. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Features Local Semantics Muhammad Anis uddin Nasir Scaling Online Social Networks 7/24
  • 16. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Features Local Semantics Load Balancing Muhammad Anis uddin Nasir Scaling Online Social Networks 7/24
  • 17. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Features Local Semantics Load Balancing Fault Tolerant Muhammad Anis uddin Nasir Scaling Online Social Networks 7/24
  • 18. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Features Local Semantics Load Balancing Fault Tolerant Dynamic Muhammad Anis uddin Nasir Scaling Online Social Networks 7/24
  • 19. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Features Local Semantics Load Balancing Fault Tolerant Dynamic Low Replication Overhead Muhammad Anis uddin Nasir Scaling Online Social Networks 7/24
  • 20. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Architecture Partition Manager Muhammad Anis uddin Nasir Scaling Online Social Networks 8/24
  • 21. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Architecture Partition Manager Directory Service Muhammad Anis uddin Nasir Scaling Online Social Networks 8/24
  • 22. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Architecture Partition Manager Directory Service Local Directory Service Muhammad Anis uddin Nasir Scaling Online Social Networks 8/24
  • 23. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Architecture Partition Manager Directory Service Local Directory Service Replication Manager Muhammad Anis uddin Nasir Scaling Online Social Networks 8/24
  • 24. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion SPAR Algorithm Heuristic Muhammad Anis uddin Nasir Scaling Online Social Networks 9/24
  • 25. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion SPAR Algorithm Heuristic Greedy Optimization Muhammad Anis uddin Nasir Scaling Online Social Networks 9/24
  • 26. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion SPAR Algorithm Heuristic Greedy Optimization Local Search Muhammad Anis uddin Nasir Scaling Online Social Networks 9/24
  • 27. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion SPAR Algorithm Heuristic Greedy Optimization Local Search Node Add/Remove Muhammad Anis uddin Nasir Scaling Online Social Networks 9/24
  • 28. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion SPAR Algorithm Heuristic Greedy Optimization Local Search Node Add/Remove Edge Add/Remove Muhammad Anis uddin Nasir Scaling Online Social Networks 9/24
  • 29. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion SPAR Algorithm Heuristic Greedy Optimization Local Search Node Add/Remove Edge Add/Remove 3 Configurations Muhammad Anis uddin Nasir Scaling Online Social Networks 9/24
  • 30. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion SPAR Algorithm Heuristic Greedy Optimization Local Search Node Add/Remove Edge Add/Remove 3 Configurations Server Add/Remove Muhammad Anis uddin Nasir Scaling Online Social Networks 9/24
  • 31. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion SPAR Algorithm Heuristic Greedy Optimization Local Search Node Add/Remove Edge Add/Remove 3 Configurations Server Add/Remove Redistribution Muhammad Anis uddin Nasir Scaling Online Social Networks 9/24
  • 32. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion SPAR Algorithm Heuristic Greedy Optimization Local Search Node Add/Remove Edge Add/Remove 3 Configurations Server Add/Remove Redistribution Let it fill Muhammad Anis uddin Nasir Scaling Online Social Networks 9/24
  • 33. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion SPAR Algorithm Heuristic Greedy Optimization Local Search Node Add/Remove Edge Add/Remove 3 Configurations Server Add/Remove Redistribution Let it fill Muhammad Anis uddin Nasir Scaling Online Social Networks 9/24
  • 34. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion SPAR Algorithm Heuristic Greedy Optimization Local Search Node Add/Remove Edge Add/Remove 3 Configurations Server Add/Remove Redistribution Let it fill Muhammad Anis uddin Nasir Scaling Online Social Networks 9/24
  • 35. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion SPAR Algorithm Heuristic Greedy Optimization Local Search Node Add/Remove Edge Add/Remove 3 Configurations Server Add/Remove Redistribution Let it fill Muhammad Anis uddin Nasir Scaling Online Social Networks 9/24
  • 36. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion SPAR Algorithm Heuristic Greedy Optimization Local Search Node Add/Remove Edge Add/Remove 3 Configurations Server Add/Remove Redistribution Let it fill Muhammad Anis uddin Nasir Scaling Online Social Networks 9/24
  • 37. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Features Distributed Partitioning Muhammad Anis uddin Nasir Scaling Online Social Networks 10/24
  • 38. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Features Distributed Partitioning k-way Partitioning Muhammad Anis uddin Nasir Scaling Online Social Networks 10/24
  • 39. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Features Distributed Partitioning k-way Partitioning Load Balancing Muhammad Anis uddin Nasir Scaling Online Social Networks 10/24
  • 40. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Features Distributed Partitioning k-way Partitioning Load Balancing Low Inter-communication Overhead Muhammad Anis uddin Nasir Scaling Online Social Networks 10/24
  • 41. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Features Distributed Partitioning k-way Partitioning Load Balancing Low Inter-communication Overhead Local Search Muhammad Anis uddin Nasir Scaling Online Social Networks 10/24
  • 42. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Overview Sampling Policies Muhammad Anis uddin Nasir Scaling Online Social Networks 11/24
  • 43. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Overview Sampling Policies Local Muhammad Anis uddin Nasir Scaling Online Social Networks 11/24
  • 44. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Overview Sampling Policies Local Random Muhammad Anis uddin Nasir Scaling Online Social Networks 11/24
  • 45. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Overview Sampling Policies Local Random Hybrid Muhammad Anis uddin Nasir Scaling Online Social Networks 11/24
  • 46. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Overview Sampling Policies Local Random Hybrid Swapping Policies Muhammad Anis uddin Nasir Scaling Online Social Networks 11/24
  • 47. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Overview Sampling Policies Local Random Hybrid Swapping Policies Energy Function Muhammad Anis uddin Nasir Scaling Online Social Networks 11/24
  • 48. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Overview Sampling Policies Local Random Hybrid Swapping Policies Energy Function Simulated Annealing Muhammad Anis uddin Nasir Scaling Online Social Networks 11/24
  • 49. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Overview Sampling Policies Local Random Hybrid Swapping Policies Energy Function Simulated Annealing Algorithm Muhammad Anis uddin Nasir Scaling Online Social Networks 11/24
  • 50. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Overview Sampling Policies Local Random Hybrid Swapping Policies Energy Function Simulated Annealing Algorithm Hybrid Sampling Muhammad Anis uddin Nasir Scaling Online Social Networks 11/24
  • 51. Motivation Algorithms SPAR Our Contribution JA-BE-JA Evaluation Conclusion Overview Sampling Policies Local Random Hybrid Swapping Policies Energy Function Simulated Annealing Algorithm Hybrid Sampling Simulated Annealing Muhammad Anis uddin Nasir Scaling Online Social Networks 11/24
  • 52. Motivation Algorithms Challenges Our Contribution Proposed Algorithm Evaluation Conclusion 1 Motivation 2 Algorithms SPAR JA-BE-JA 3 Our Contribution Challenges Proposed Algorithm 4 Evaluation Datasets Implementation Results 5 Conclusion Muhammad Anis uddin Nasir Scaling Online Social Networks 12/24
  • 53. Motivation Algorithms Challenges Our Contribution Proposed Algorithm Evaluation Conclusion Challenges Global View Muhammad Anis uddin Nasir Scaling Online Social Networks 13/24
  • 54. Motivation Algorithms Challenges Our Contribution Proposed Algorithm Evaluation Conclusion Challenges Global View Partition Manager Muhammad Anis uddin Nasir Scaling Online Social Networks 13/24
  • 55. Motivation Algorithms Challenges Our Contribution Proposed Algorithm Evaluation Conclusion Challenges Global View Partition Manager Replication Overhead Muhammad Anis uddin Nasir Scaling Online Social Networks 13/24
  • 56. Motivation Algorithms Challenges Our Contribution Proposed Algorithm Evaluation Conclusion Challenges Global View Partition Manager Replication Overhead Load Balancing Muhammad Anis uddin Nasir Scaling Online Social Networks 13/24
  • 57. Motivation Algorithms Challenges Our Contribution Proposed Algorithm Evaluation Conclusion Proposed Algorithm SPAR + JA-BE-JA Muhammad Anis uddin Nasir Scaling Online Social Networks 14/24
  • 58. Motivation Algorithms Challenges Our Contribution Proposed Algorithm Evaluation Conclusion Proposed Algorithm SPAR + JA-BE-JA Gossip Learning Muhammad Anis uddin Nasir Scaling Online Social Networks 14/24
  • 59. Motivation Algorithms Challenges Our Contribution Proposed Algorithm Evaluation Conclusion Proposed Algorithm SPAR + JA-BE-JA Gossip Learning Simulated Annealing Muhammad Anis uddin Nasir Scaling Online Social Networks 14/24
  • 60. Motivation Algorithms Challenges Our Contribution Proposed Algorithm Evaluation Conclusion Proposed Algorithm SPAR + JA-BE-JA Gossip Learning Simulated Annealing Optimal Replication Muhammad Anis uddin Nasir Scaling Online Social Networks 14/24
  • 61. Motivation Algorithms Datasets Our Contribution Implementation Evaluation Results Conclusion 1 Motivation 2 Algorithms SPAR JA-BE-JA 3 Our Contribution Challenges Proposed Algorithm 4 Evaluation Datasets Implementation Results 5 Conclusion Muhammad Anis uddin Nasir Scaling Online Social Networks 15/24
  • 62. Motivation Algorithms Datasets Our Contribution Implementation Evaluation Results Conclusion Datasets Synthetic Graphs 0 http://snap.stanford.edu/data/ Muhammad Anis uddin Nasir Scaling Online Social Networks 16/24
  • 63. Motivation Algorithms Datasets Our Contribution Implementation Evaluation Results Conclusion Datasets Synthetic Graphs Facebook Graphs 0 http://snap.stanford.edu/data/ Muhammad Anis uddin Nasir Scaling Online Social Networks 16/24
  • 64. Motivation Algorithms Datasets Our Contribution Implementation Evaluation Results Conclusion Datasets Synthetic Graphs Randomized Facebook Graphs 0 https://gephi.org/ Muhammad Anis uddin Nasir Scaling Online Social Networks 17/24
  • 65. Motivation Algorithms Datasets Our Contribution Implementation Evaluation Results Conclusion Datasets Synthetic Graphs Randomized Clustered Facebook Graphs 0 https://gephi.org/ Muhammad Anis uddin Nasir Scaling Online Social Networks 17/24
  • 66. Motivation Algorithms Datasets Our Contribution Implementation Evaluation Results Conclusion Datasets Synthetic Graphs Randomized Clustered Highly Clustered Facebook Graphs 0 https://gephi.org/ Muhammad Anis uddin Nasir Scaling Online Social Networks 17/24
  • 67. Motivation Algorithms Datasets Our Contribution Implementation Evaluation Results Conclusion Datasets Synthetic Graphs Randomized Clustered Highly Clustered Facebook Graphs 150 nodes, 3386 edges 0 https://gephi.org/ Muhammad Anis uddin Nasir Scaling Online Social Networks 17/24
  • 68. Motivation Algorithms Datasets Our Contribution Implementation Evaluation Results Conclusion Datasets Synthetic Graphs Randomized Clustered Highly Clustered Facebook Graphs 150 nodes, 3386 edges 224 nodes, 6384 edges 0 https://gephi.org/ Muhammad Anis uddin Nasir Scaling Online Social Networks 17/24
  • 69. Motivation Algorithms Datasets Our Contribution Implementation Evaluation Results Conclusion Datasets Synthetic Graphs Randomized Clustered Highly Clustered Facebook Graphs 150 nodes, 3386 edges 224 nodes, 6384 edges 786 nodes, 60050 edges 0 https://gephi.org/ Muhammad Anis uddin Nasir Scaling Online Social Networks 17/24
  • 70. Motivation Algorithms Datasets Our Contribution Implementation Evaluation Results Conclusion Implementation SPAR Proposed Algorithm Metric Replication Overhead Muhammad Anis uddin Nasir Scaling Online Social Networks 18/24
  • 71. Motivation Algorithms Datasets Our Contribution Implementation Evaluation Results Conclusion Evaluation of Replication Overhead Replication Factor Fault tolerance replicas reduce replication overhead Proposed Algorithm performs better than SPAR 0 replication overhead = number of replicas/number of users Muhammad Anis uddin Nasir Scaling Online Social Networks 19/24
  • 72. Motivation Algorithms Datasets Our Contribution Implementation Evaluation Results Conclusion Evaluation of Replication Overhead Number of Servers Less Replication overhead in the case of proposed algorithm Proposed Algorithm performs better in the case of high clusterization 0 replication overhead = number of replicas/number of users Muhammad Anis uddin Nasir Scaling Online Social Networks 20/24
  • 73. Motivation Algorithms Datasets Our Contribution Implementation Evaluation Results Conclusion Evaluation of Replication Overhead Number of Servers Less Replication overhead in the case of proposed algorithm Proposed Algorithm performs better in case of high clusterization 0 replication overhead = number of replicas/number of users Muhammad Anis uddin Nasir Scaling Online Social Networks 21/24
  • 74. Motivation Algorithms Our Contribution Evaluation Conclusion 1 Motivation 2 Algorithms SPAR JA-BE-JA 3 Our Contribution Challenges Proposed Algorithm 4 Evaluation Datasets Implementation Results 5 Conclusion Muhammad Anis uddin Nasir Scaling Online Social Networks 22/24
  • 75. Motivation Algorithms Our Contribution Evaluation Conclusion Conclusion Distributed social-based partitioning Muhammad Anis uddin Nasir Scaling Online Social Networks 23/24
  • 76. Motivation Algorithms Our Contribution Evaluation Conclusion Conclusion Distributed social-based partitioning Local Semantics Muhammad Anis uddin Nasir Scaling Online Social Networks 23/24
  • 77. Motivation Algorithms Our Contribution Evaluation Conclusion Conclusion Distributed social-based partitioning Local Semantics Reduced Replication overhead compared to SPAR Muhammad Anis uddin Nasir Scaling Online Social Networks 23/24
  • 78. Motivation Algorithms Our Contribution Evaluation Conclusion Conclusion Distributed social-based partitioning Local Semantics Reduced Replication overhead compared to SPAR Better load balancing using k-way partitioning Muhammad Anis uddin Nasir Scaling Online Social Networks 23/24
  • 79. Motivation Algorithms Our Contribution Evaluation Conclusion Conclusion Distributed social-based partitioning Local Semantics Reduced Replication overhead compared to SPAR Better load balancing using k-way partitioning Transparent Scaling Muhammad Anis uddin Nasir Scaling Online Social Networks 23/24
  • 80. Motivation Algorithms Our Contribution Evaluation Conclusion Scaling Online Social Networks: extended SPAR using Gossip Learning Presented by: Muhammad Anis uddin Nasir Coworker: Maria Stylianou Supervised by: Sarunas Girdzijauskas KTH Royal Institute of Technology December 5, 2012 Muhammad Anis uddin Nasir Scaling Online Social Networks 24/24