Lab seminar Towards A Maximum-Flow-Based Service Composition (for Multiple & Concurrent Service Composition) Han, Sang Woo Networked Media Lab. Dept. of Information and Communications Gwangju Institute of Science and Technology
Contents Ph.D. Research Topics Introduction Motivation Related Work Research Outline Proposed Service Composition Scheme System Model & Problem Statement Problem Solving Methods Discussion Summary
Workflow-driven Control and Management Framework for Dynamic Service Composition Hierarchical Abstraction Structure for Programmable Network and Computing Environments Workflow-driven Dynamic Service Composition Capability-based Service Matchmaking and Negotiation Ph.D. Research Topics
Mobile multimedia services Live media streaming Personalized internet broadcasting Multi-party video conferencing Full Web Browsing Challenges QoS support between devices having heterogeneous network & device capability Live Content Sharing over Mobile P2P Networks Mobile P2P Networks Your ContentYour FriendsYour Device Media Consumers Media Producers capability gap QoS-aware service composition
BCP (bounded composition probing protocol) Hop-by-hop probing processing & optimal composition selection Not supporting multiple composition in same time [HPDC 04] Spidernet: An integrated peer-to- peer service composition framework
SeSCo (seamless service composition) Hierarchical service overlay network configuration Discovery + matching + coordination MM 05] Seamless Service Composition (SeSCo) in Pervasive Environments
Goal Multiple & concurrent service composition (modeling) Challenges Existing schemes does not consider multiple & concurrent service composition Thus, next composition requests have to be blocked in processing a composition job composition processing time become longer! Approach Casting the composition problem into maximum flow network problem Multiple sources, multiple sinks Possible maximum flow out of certain sources or into all sinks Expected Result Automated Service Composition Graph (in Polynomial-Time) Research Outline
Media-Service-Oriented Virtualized Computing & Networking Testbed Networked Cameras Storage service Telecommunication service Video producing service Web servers Replica facilities Content servers Encoding, transcoding, and decoding services Presence service
Use Case 1) request for interactive broadcasting Apps portal 2) posting & announcement 3) application-on-demand 5) quotation 6) reservation & payment 8) commit Transcoding service Video scaling service Text embedding service Multicast connector service network services offered by service providers 4) query & negotiation Application #1 7) service path reservation & payment Application #2 Application #3 interactive & personalized broadcasting users 4K cinema video conferencing content providers Service path 1 … multimedia mashup
Preliminary System Model ApplicationTestbed Topology Input: Multiple applications and testbed topology Output: The graphs of service composition for the applications
(DHT-based) Service Discovery Service Instantiating (according to # of apps) Step 1. Service Finding
Step 2. Configuring Network Unit Capacity Maximum Flow Network
Step 3. Service Paths Finding
Input: Graph G with flow capacity c, a source node s, and a sink node t Output: A flow f from s to t which is a maximum 1. f(u,v) 0 for all edges (u,v) 2. While there is a path p from s to t in G f, such that c f (u,v)>0 for all edges (u,v) ∈ p: 1. f(u,v) f(u,v) + c f (p) 2. f(v,u) f(v,u) – c f (p) Service Path Finding Using Maximum Flow Algorithm Ford-Fulkerson Algorithm
How to evaluate? To measure service composition processing time per application in large-scale virtualized computing & networking testbed Need more criteria… Network capacities consideration System model update using weighted maximum flow algorithm Adaptive composition Feedback-driven resource/service adaptation Stabilization in dynamic situation Load balancing Complex application design Workflow-pattern-based specification Discussion
Summary Preliminary system model for multiple & concurrent service composition Service composition approach based on network optimization method Haven’t I done an evaluation yet.
J. Jin and K. Nahrstedt, “Source-based QoS Service Routing in Distributed Service Networks,” in Proc. ICC, Paris, France, N. J.A. Harvey, R. E. Ladner, L. Lovász, and T. Tamir, “Semi-matchings for Bipartite Graphs and Load Balancing,” Algorithms and Data Structures, L. R. Ford, and D. R. Fulkerson, “Solving the Transportation Problem,” Management Science, Vol. 3, pp S. Kalasapur, M. Kumar, and B. Shirazi, “Seamless service composition ( SeSCo) in pervasive environments,” in Proc. ACM int’l workshop on Mult imedia Service Composition, New York, NY, X. Gu and K. Nahrstedt, “Distributed Multimedia Service Composition with Statistical QoS Assurances,” IEEE Trans. on Multimedia, Vol. 8, No. 1, Feb References