Replica Placement Heuristics of Application-level Multicast Chia-Hsing Yu Jiahua He CSE of UCSD
Project Presentation of CSE 222A Outline Multicast and RMX Model and Heuristics Simulation and Results Conclusion and Future Work 2019/4/8 Project Presentation of CSE 222A
Application-level Multicast Goal Distribute Contents to Many Clients Problem How to reduce the load of the central server? How to reduce the response time of requests? Replication at different servers 2019/4/8 Project Presentation of CSE 222A
RMX: Reliable Multicast proXy TCP SRM: Reliable IP Multicast 2019/4/8 Project Presentation of CSE 222A
Project Presentation of CSE 222A RMX Semantic reliability information representation of information Sender can lower the stream resolution if the network load is heavy 2019/4/8 Project Presentation of CSE 222A
Project Presentation of CSE 222A Existing Problems Only sources, no replicas No request, only recovery request Static RMXs in network Static configuration of data groups 2019/4/8 Project Presentation of CSE 222A
Project Presentation of CSE 222A Related works Replication in unstructured P2P (Princeton) Owner, Path, Random PAST(Microsoft and Rice) Nodes with similar id’s OceanStore (Berkeley) On or near the clients Focus on persistent storage with versions Chain (Cornell) Machines with replicas of a same file form a chain Focus on availability 2019/4/8 Project Presentation of CSE 222A
Project Presentation of CSE 222A Model and Heuristics Fixed sources and dynamic replicas Streaming multicast on demand No replication Baseline Replication on path FIFO LRU Color 2019/4/8 Project Presentation of CSE 222A
Project Presentation of CSE 222A Baseline Only sources, no replicas Learning bridge scheme to search Learn routing information from incoming data Soft state: periodically refresh Request suppression Ideal condition: no loss 2019/4/8 Project Presentation of CSE 222A
Project Presentation of CSE 222A FIFO and LRU Replication on path Broadcast to search FIFO: Remove the oldest one if no space LRU: Order the files by last usage 2019/4/8 Project Presentation of CSE 222A
Project Presentation of CSE 222A Color Graph coloring Neighbors with different colors (files) from mine Can get more different files from neighbors Remove the file with nearest replica Visiting Frequency More frequently visited, more possible to be visited Cost function: dist * freq dist: distance to the nearest replica freq: visiting frequency Upper bound of the cost if removed 2019/4/8 Project Presentation of CSE 222A
Project Presentation of CSE 222A Simulator Event-driven Simulator New Event New Event New Event Event Handler Min Heap Earliest Event 2019/4/8 Project Presentation of CSE 222A
Project Presentation of CSE 222A Simulator(2) Stream-level Simulation SIM_SEND_STREAM( bit rate, length ) Input Network Topology Host Resources Stream Sources User Requests 2019/4/8 Project Presentation of CSE 222A
Experiment Configuration Network Topology Binary Tree Host Resources 127 hosts (data groups) Hard disk size variable Stream Sources 1270 sources (average 10 per host) 500 Kbps, 8000 seconds each Randomly distributed User Request Total number variable Experiment Span 100 hours 2019/4/8 Project Presentation of CSE 222A
Experiment Configuration (2) Variances Number of requests: 211 ~ 218 Hard disk size: 8G ~ 128G Metrics Client view average response time Server view load (number of streams per RMX) load balance (standard deviation of load) System view throughput 2019/4/8 Project Presentation of CSE 222A
Client View Avg. Response Time vs. # of Requests About 30% improvement 2019/4/8 Project Presentation of CSE 222A
Client View Avg. Response Time vs. Disk Size Disk size outperforms replication strategy 2019/4/8 Project Presentation of CSE 222A
Server View Avg. # of Streams vs. # of Requests About 50% improvement 2019/4/8 Project Presentation of CSE 222A
Server View S.D. # of Streams vs. # of Requests About 50% improvement 2019/4/8 Project Presentation of CSE 222A
Server View Avg. # of Streams vs. Disk Size Disk size outperforms replication strategy 2019/4/8 Project Presentation of CSE 222A
Server View S.D. # of Streams vs. Disk Size Disk size outperforms replication strategy 2019/4/8 Project Presentation of CSE 222A
System View Throughput vs. # Requests About 25% improvement 2019/4/8 Project Presentation of CSE 222A
System View Throughput vs. Disk Size Upper bound 25.4398 System View Throughput vs. Disk Size 2019/4/8 Project Presentation of CSE 222A
Project Presentation of CSE 222A Contributions Implement and analyze Baseline, FIFO, LRU algs Propose and verify Color heuristics Avg. response time: up to 30% improvement Load: up to 50% improvement Load balance: up to 50% improvement Throughput: up to 25% improvement 2019/4/8 Project Presentation of CSE 222A
Project Presentation of CSE 222A Future Works Biased requests Heterogeneous environment (hosts, links, streams) Random forward More sophisticated heuristics Experiment in real environment 2019/4/8 Project Presentation of CSE 222A