Thesis Presentation P2 P Vo D On Internet Rodrigo Godoi

Loading...

Flash Player 9 (or above) is needed to view presentations.
We have detected that you do not have it on your computer. To install it, go here.

0 comments

Post a comment

    Post a comment
    Embed Video
    Edit your comment Cancel

    Favorites, Groups & Events

    Thesis Presentation P2 P Vo D On Internet Rodrigo Godoi - Presentation Transcript

    1. UniversitatAutònoma de Barcelona
      Computer Architecture & Operating Systems Department
      P2P-VoD on Internet: Fault Tolerance and Control Architecture
      Rodrigo Godoi
      Advisor: Dr. Porfidio Hernández Budé
      Barcelona, July 2009.
    2. Contents
    3. Contents
    4. Video on Demand - VoD
      • Multimedia service
      • Asynchronous requests
      • Every client enjoys entire content
      • Long sessions (> 60 min.)
    5. VoD - requirements and constraints
      Scalability
      Large scale Video on Demand - LVoD
      • Clients: thousands and disperse
      • Multimedia contents: huge catalogue
      Multicast
      Peer-to-Peer
      Internet
      Soft real-time
      • Time limit on handling data
      • Quality of Service (QoS)
      Control Architecture
      Fault Tolerance
    6. Multicast
      - implementations
      IP Multicast:
      • Source tree (e.g. PIM-DM)
      • Shared tree (e.g. PIM-SM)
      Application Layer Multicast ALM
      (e.g. NICE, ALMI)
      Overlay Multicast
      (e.g. OMNI)
    7. Multicast- Patching
      Patching: multicast technique for multimedia data delivery
      Base stream
      t= 0
      t = 6
      Multicast
      Unicast
      Patchstream
    8. Peer-to-Peer
      Free cooperation of equals in view of the performance of a common task.
      • Takes advantage of resources (storage, cycles, content, human presence) available at the edges of the Internet.
      • Usage: file sharing, distributed multimedia systems, high performance computing.
      Synchronised usage of peers resources: Collaboration Groups
    9. Tracker
      Peers
      Peers
      Peers
      Supernodes
      Peer-to-Peer - classification
      Peers
      Location mechanism
    10. Peer-to-Peer - classification
      Overlay topology
      Chain
      Mesh
      Tree
    11. Internet environment
      • Worldwide scale
      • Heterogeneous environment
      • Best-effort service
      • Exponential growth rate
      Organisation:
      • Autonomous Systems (AS): collection of connected IP routing prefixes under the control of one or more network operators (ISPs, universities, companies)
      • Network arranged by dimension and purpose (LAN, WAN, MAN)
      • Modeled by complex network theory
      Clustering coefficient
      Average path length
    12. Problem
      Large scale
      system
      Failures
      Network
      Server
      Peers
      Frequent arrivals/departures
    13. Problem
      Failures/Errors treatment
      Input rate fluctuation
      Source crash
      Cushion buffer
      QoS
      Start-up delay
      VoD service must…
      • respect deadlines
      • provide low start-up delay
      • have a clever buffer usage
      • enforce low control overhead
      Control Architecture
      Fault Tolerance
    14. Control relevance
      P2P and Multicast
      Heterogeneity: Internet, peers capabilities, lifetimes
      Resources sharing
      Delivery Architecture
      Control Architecture
    15. FaultTolerance
      Consequence of a failure
      System defect
      Do not solve the fault
    16. State of the art
    17. Contents
    18. Goal of the Thesis
      To assess Control impact and propose a Fault Tolerance Scheme for P2P-VoD service on the Internet.
      • Scalability
      • Flexibility
      • Reliability
      • Efficiency
      • Low overhead
      QoS
    19. Contents
    20. System architecture
      Clients overlay topology
      Distributed proxy servers
      Distributed video servers
      Internet
      Clients
      P2P Collaborations
      Servers overlaytopology
      IP Multicast zone
      Internet Autonomous System
    21. The Failure Management Process
      Basis of Fault Tolerance Mechanisms
      • Income stream monitoring
      • Heartbeat messages
      • Centralised
      • Subsequent queries
      Detection
      Recovery
      Maintenance
      • Network infrastructure
      • Peer status
    22. Load and Time metrics
      Load cost
      Volume of control messages that flows through the system on failure management processes
      Control overhead - congestion, bandwidth consumption
      Time cost
      Time consumed by solving peer failures
      Control efficiency - start-up delay, buffer usage
    23. Background: VoD service schemes
      Gather different aspects of P2P-VoD services
    24. PCM/MCDB
      PCM
      MCDB
      PCM: Patch Collaboration Manager
      MCDB: Multicast Channel Distributed Branching
      Bypass
    25. FaultTolerance - PCM/MCDB
      MCDB
      Maintenance messages
      Ch.M0
      Ch.M1
      Recovery messages
      Ch.M2
      • Centralised recovery.
      • IP Multicast tree rearrangement
      Detection messages
    26. P2Cast
      Base Stream
      Patch Stream
      • Clients are divided into sessions according to the arrival time in the system (session threshold parameter - T)
      • Best-fit algorithm: Peer with great amount of available bandwidth is selected as parent
      VoD Server
      Session 4
      Session 3
      31.0
      20.0
      34.0
      35.0
      24.0
      21.0
      37.0
      35.8
      39.9
      35.5
      27.0
      40.0
      T
    27. FaultTolerance - P2Cast
      • Source peers (Parents) failures provoke stream disruption
      • Subsequent recovery queries
      Detection messages
      VoD Server
      Recovery messages
      Session 4
      Session 3
      31.0
      20.0
      34.0
      35.0
      24.0
      21.0
      37.0
      35.8
      39.9
      35.5
      27.0
      40.0
    28. Load cost
      Control messages
      PCM/MCDB
      P2Cast
      Heartbeats
      Heartbeats
      Detection
      IP multicast rearrangement
      Recovery request
      Subsequent queries
      Recovery
      Peers status
      Routers status
      Routers status
      Maintenance
    29. Time cost
      Time consumption
      Recovery messages
      Detection
      Path (network theory): small-world effect
      latency
      PCM/MCDB
      Subsequent queries
      Recovery messages
      Detection
      Path (network theory): small-world effect
      latency
      P2Cast
    30. Contents
    31. TheFaultToleranceScheme (FTS)
      Cushion
      Delivery
      Collaboration
      Altruist
      FTS stands on peers capabilities:
      • Input / Output bandwidth
      • Buffer size
      bwo
      bwi
      buffer
      8
      9
      10
      11
      12
      L
      6
      7
      13
      Fault Tolerance Groups
      Buffer In
      Buffer Out
    32. TheFaultToleranceScheme (FTS)
      MN
      C1
      C2
      Cushion
      Delivery
      Gen. purpose
      Altruist
      MN
      Fault Tolerance Groups
      MN
      MN
      L
      1
      2
      3
      1
      2
      3
      t = 0
      L
      4
      5
      1
      2
      3
      7
      8
      9
      4
      5
      6
      1
      2
      3
      10
      7
      MN
      13
      14
      L
      15
      10
      11
      12
      8
      9
      4
      5
      1
      2
      3
      t = 0
      t = 0
      C1
      7
      L
      4
      5
      6
      1
      2
      3
      6
      7
      C1
      10
      11
      L
      12
      7
      8
      9
      4
      5
      6
      9
      10
      6
      7
      8
      13
      14
      Video
      8
      9
      10
      6
      7
      13
      14
      15
      11
      12
      3
      4
      5
      1
      2
      Manager Node
      MN
      C2
      7
      L
      4
      5
      6
      1
      2
      3
      14
      15
      11
      12
      13
      t = 3
      t = 3
      t = 10
      C1
      Video
      8
      9
      10
      6
      7
      13
      14
      15
      11
      12
      3
      4
      5
      1
      2
      7
      Video
      8
      9
      10
      6
      13
      14
      15
      11
      12
      3
      4
      5
      1
      2
      FTS Collaborators
      [t = 17]
    33. Load and Time costswiththe FTS
      The proposed Fault Tolerance Scheme…
      • distributes the control through Manager Nodes
      • removes subsequent queries during recovery
      • eliminates messages for peers status maintenance
      • can detect failures through heartbeats (FTS I) and income stream monitoring (FTS II)
    34. Contents
    35. Simulation tool: VoDSim
      Computational simulations provide a more dynamic and scalable analysis
      • Discrete event-driven model
      • More than 50 classes in C++
      • Over 46.000 lines
      • Peer arrival rate: Poisson
      • Content popularity: Zipf
    36. VoDSim extensions
      • Implementation of ALM service scheme: P2Cast
      • Peers disruptions: Weibull
      • FMP instrumentation:
      Load and Time costs measurement
      Fault probability
      Lifetime
    37. Contents
    38. Experimental Results
    39. Failure Management Process validation
    40. Control vs. Multimedia traffic
      Analytical results (PCM/MCDB and P2Cast)
      ∆w = 13%-39%
      ∆w =13%- 37%
      ∆w = 10%-28%
      Simulated results (P2Cast)
    41. Load cost analysis
    42. Load cost analysis
    43. Time cost analysis
      High latency
      Time cost increment
      Recovery control messages
    44. Time cost analysis
      Download rate: 1500 kb/s (750+750)
      Cushion buffer
      56MB
      11MB
      1 min.
      5 min.
      Start-up delay
    45. Experimental Results
    46. Load cost analysis
      Cost increment
      Cost reduction
      FTS I - heartbeat detection
      FTS II - buffer monitoring detection
    47. Load cost analysis
      • Overhead reduction
      • Scalability
    48. Time cost analysis
      High latency
      Time cost increment
      Volume of communication
      FTS
      • Efficiency
      Cushion buffer
      56MB
      11MB
      Start-up delay
      5 min.
      1 min.
      τ = 1/(2·fHB)
      FTS I - heartbeat detection
      FTS II - buffer monitoring detection
    49. Fault Tolerance service performance
      • Reliability
      • Flexibility
      Altruist buffer 338MB
      Altruist buffer 102MB
    50. Contents
    51. Conclusions
      Control mechanism plays a crucial role on designing P2P-VoD systems
      Load cost
      Control overhead: network congestion, bandwidth resources
      Time cost
      Efficiency: buffer usage, start-up delay
      Load and Time costs trade-off
      Reduction of Load and Time costs
      Quality of Service
    52. Conclusions
      The Fault Tolerance Scheme…
      • is flexible for Internet use
      • presents hierarchical control structure
      • has scalable backup mechanism
      • do not demand extra data communication and dedicated resources
      • is able to guarantee system reliability
      • reduces Load and Time costs
    53. Contents
    54. Future Work
      • Application and assessment of the FTS in a wide range of VoD architectures and service policies
      • Implementation of the FTS in a simulation environment
      • FTS improvement: storing parts of non-visualized contents; using non-volatile storage devices (e.g. Solid State Disk drives)
      • Addition of VCR / DVD-like operations
      • Usage of clients behaviour information to improve system performance
    55. P2P-VoD on Internet: Fault Tolerance and Control Architecture
      Rodrigo Godoi
      Thankyou
      Gracias
      Obrigado
      Barcelona, July 2009.
    56. TheFaultToleranceScheme (FTS)
      Server
      Fault Tolerance Group
      Control comm.
      Manager Node
      Clients
      FTG members
      Architecture elements
      • Server: content seed
      • Peer: multimedia client / source
      • FTG member: collaborator in the FTS
      • Distributed backup: flexibility and reliability.
      • Built on the fly: backup do not need retransmission.
      • P2P based: mechanism uses own system available resources.
      • Hierarchical control: scalability and deployment.
      • Manager Node: organize and monitor FTG
    57. The FTS formationlaw
      While
      If
      then
      Add Collaborator to FTG
      If
      Peers’ bandwidth greater than playback rate (bw≥Vpr)
      then
      New FTG.
      Input parameters
      FTG size
      Distributed backup
      Service conditions
    58. The FTS formationlaw
      While
      MN
      C1
      C2
      C3
      A · 2
      A · 3
      A · 5
      A · 5
      15
      15
      15
      15
      If
      then
      Add Collaborator to FTG.
      If
      if:
      Peers’ bandwidth lower than playback rate (bw<Vpr)
      then
      New FTG.
      if:
      A
      Video
      C
      B
      D
      E
      H
      G
      F

      Collaboration Capacity
      A
      Buffer and Bandwidth constraints

      C1
      A*
      B*

      A*
      B*
      C2
      FTG size
      MN [500kb/s]
      C1 [200kb/s]
      C2 [300kb/s]
      C3 [500kb/s]
      Vpr [1500kb/s]

      A*
      B*
      C3
      MN
      A*

      B*
    59. TheFaultToleranceScheme (FTS)
      MN
      C1
      C2
      Client
      MN
      I
      II
      III
      IV
      Creation of Fault Tolerance Groups
      Local Server
      Collab. availability
      FTS ack.
      Join to FTG
      Start new FTG and become Manager Node
      Standby status
    60. TheFaultToleranceScheme (FTS)
      Standby Peer
      I
      II
      FTG: complexity and maintenance
      O(NCFTG)
      MN
      MN
      failure
      Local Server
      Member failure
      C1
      C1
      C2
      Designation of new MN
      Restoring the FTG
      Restoring the FTG
    61. Evaluation environment
      Underlying network: GT-ITM topology generator
      Transit-stub model
      • 1Transit domain (3 routers)
      • 6 Stub domain (54 routers)
      Service schemes
      • ALM / tree-based P2P (P2Cast)
      • IP Multicast / mesh-based P2P (PCM/MCDB)
      Network protocols
      • Unicast: OSPF
      • IP Multicast: IGMP and PIM-SM
    62. Evaluation environment
      Network Protocols
    63. Conclusions - publications
    SlideShare Zeitgeist 2009

    + Rodrigo GodoiRodrigo Godoi Nominate

    custom

    75 views, 0 favs, 0 embeds more stats

    More info about this document

    © All Rights Reserved

    Go to text version

    • Total Views 75
      • 75 on SlideShare
      • 0 from embeds
    • Comments 0
    • Favorites 0
    • Downloads 1
    Most viewed embeds

    more

    All embeds

    less

    Flagged as inappropriate Flag as inappropriate
    Flag as inappropriate

    Select your reason for flagging this presentation as inappropriate. If needed, use the feedback form to let us know more details.

    Cancel
    File a copyright complaint
    Having problems? Go to our helpdesk?

    Categories