Availability Simulation of Peer-to-Peer Architectural Styles Simon - - PowerPoint PPT Presentation

Availability Simulation

• f Peer-to-Peer Architectural Styles

Simon Giesecke, Timo Warns, Wilhelm Hasselbring

Referee: Timo Warns

2

Availability Simulation of Peer-to-Peer Architectural Styles

Motivation

 Evaluation of availability of P2P services  Specifics of P2P context impacting availability

 Failure distribution of peers  Means of handling failures  Dynamic architecture / topology

 How to integrate these aspects?

 Focus: Architectural Style

3

Availability Simulation of Peer-to-Peer Architectural Styles

Approach

 Conceptual framework

 P2P styles  P2P architectures  P2P systems

 Evaluation by simulation

 “most real-world systems are

too complex to allow realistic models to be evaluated analytically”

Law and Kelton, 2000

 Flexible Architectural Style Architecture Simulated System Real-World System Fault Characteristics Predicted Availability Actual Availability

4

Availability Simulation of Peer-to-Peer Architectural Styles

Peer-to-Peer Styles

 Classification scheme

 Type of decentralization

 Decentralized, hybrid, super-peer

 Type of communication

 Direct, Indirect, Mediated

 Structural Characteristics

 Ring, Tree, Small-World Network

 Rules for evolution

 Joining / leaving of peers

 No formalisation yet

Server Peer Peer Peer Peer Super Peer Super Peer Super Peer Peer Peer Peer Peer Peer Peer Peer

5

Availability Simulation of Peer-to-Peer Architectural Styles

Architecture Description Model

 Graph-based formalism A = (N, C, ν, λ, τ)  N, C – Sets of nodes and connections  ν: C →{{n1, n2} | n1 ≠ n2 and n1, n2 in N} – Node function  λ: N → L – Labelling function  L is a set of node labels (e.g., “Peer”, “Server”, ...)  τ: T → NCT – Time mapping  τ describes evolution over time  E.g., peer p participates at system from tn to tm

=> p is in image of τ for t in [tn, tm[

6

Availability Simulation of Peer-to-Peer Architectural Styles [t2, t3[ [t0, t1[

Example Description Model

 N = {p1, ..., p4}  C = {c1, ..., c5}  λ(n) = Peer for all n in N  ν:

c ν(c)

c1 {p1, p3} c2 {p1, p2} c3 {p2, p3} c4 {p3, p4} c5 {p1, p4}

Peer Peer Peer

p1 p2 p3 c1 c2 c3

Peer Peer Peer

p1 p2 p3 c1 c2 c3

Peer

c4 p4

Peer Peer Peer

p1 p2 p3 c1 c2 c3

Peer

c4 p4 c5 [t1, t2[

 τ:

T NCT

[t0, t1[ p1, ..., p3, c1, ..., c3 [t1, t2[ p1, ..., p4, c1, ..., c4 [t2, t3[ p1, ..., p4, c1, ..., c5

7

Availability Simulation of Peer-to-Peer Architectural Styles

Simulation

 Prototype of simulator

 Based on graph formalism

 Peer model

 Derived from real-world

system

 Enhanced by classic

replication strategies

 Evaluation of availability of

replicated resources

1 2 3 4 5 6 7 8 9 10 11 12

• 0,25

0,00 0,25 0,50 0,75 1,00 1,25 1,50 1,75 2,00 2,25 2,50 2,75 3,00 3,25 3,50 3,75 4,00 4,25 4,50 ROWA Majority Consensus Scenario Class Relative Change in %

8

Availability Simulation of Peer-to-Peer Architectural Styles

Conclusions

 Conceptual framework

 Evaluation of availability of P2P services  Architectural styles, architectures, systems

 Classification scheme for architectural styles  Description model for P2P architectures  Simulator prototype

9

Availability Simulation of Peer-to-Peer Architectural Styles

Future Work

 Formalisation of architectural styles

 Graph grammars?  Benefit: Automated creation of architectures

 Formalisation of peer model

 Add peer model to input for simulation  UML?

 Development of improved simulator

 Prototype used manually created architectures and

• ne fixed peer model