Implementation of Host-based Overlay Multicast in Support of Web Based Services for RT-DVS Dennis Moen, Mark Pullen & Fei Zhao George Mason University {dmoen,mpullen,fzhao}@gmu.edu
Network Service Requirements for Real Time Distributed Virtual Simulation � Network Quality of Service (QoS) � end-to-end capacity, latency, jitter, and packet loss in a statistical sense � Multicast � many-to-many group communication � Reliable Multicast Transport � high confidence of delivery � End-to-end network status and performance monitoring � need to know what the network is doing for you � Multi-sensor systems � must manage streaming data with low latency
Internet Multicast Services Today • IP multicast over the Internet not widely deployed • IETF initial focus is on one-to-many multicast • Commercial viability lacking for IP multicast in the Internet • Result: interest in multicast based on end systems not network – End-to-end argument: push complexity up the stack – Example: TCP is complex, IP is simple
Overlay Multicast Tree E B D INTERNET INTERNET A F E G B C A D F C IP Multicast tree: G H J
XOM Overlay XOM 1 XOM 2 Router Router A B XOM 4 XOM 3 XOM 5 Router Router D C XOM 6 XOM 7
XOM Layers Generic Class Definition Interface (SRMP Example) Registry Join/leave Routing Table Group Management Security Address Routing Capacity/latency Path Management Node Demand Path Optimization Packet Send/Receive Listen to Class QoS/ Distribute Messages Ports Queueing UDP IP
XOM Group Membership B C G 2 = { B, C, D} D XOM 3 Internet XOM 1 A XOM 2 G 1 = {A, B} A B B C G 2 = { B, C, D} G 1 = {A, B} D Application B sending implies routing to group G 3 = { G 1 ? G 2 }
Group Aggregation Overlay (Optimum Path Overlay) Multicast Groups Aggregate Trees Group Members Tree Tree Links (arcs) g 0 XOM 1,2,3,4 T 0 1-4, 4-2, 4-3 g 1 XOM 1,2,3,4 g 2 XOM 1,2,3 g 3 XOM 1,2 Groups g 0 , g 1 , g 2 , g 3 share one T 0 aggregate tree T 0 . T 0 is a perfect XOM 1 match for g 0 and g 1 , but is a leaky (g 0 , g 1 , g 2 , g 3 ) match for g 2 and g 3 . Trades off path utilization inefficiency for XOM 4 Internet lower path management (g 0 , g 1 ) overhead. XOM 2 (g 0 , g 1 , g 2 , g 3 ) XOM 3 (g 0 , g 1 , g 2 )
Overlay Routing Constraints End-to-End Latency Path Constraint Access capacity Minimum tree (Rate Control) Demand Constraint Optimum Path Internet XOM XOM Simulation Application Traffic Load
XOM Functional Model Prototype Test Scenario Traffic Generator SRMP Packet Receiver Packet Sender Host Channel Routing Abstraction for Table for Multicast Channel (S,G) Channel (S,G) UDP IP
XOM Prototype Other Registries XOM Sites Internet incoming outgoing routing WAN info stats Statistics MulticastRouter* Routing LAN (Java or C++) stats routing data table multicast to/from WAN *All modules except IPmc Host IPmc Host Router are Java
XOM Lab Test Scenarios XOM 1 XOM 1 XOM 4 XOM 4 XOM 2 XOM 3 XOM XOM 2 3 Test 1. XOM n -degree of 3 Test 2. XOM n -degree of 2
Message Delay 90. 80. 2-degree 70. 3-degree 60. Delay (msec) 50. 40. 30. 20. 10. 0.0 0 500 1000 1500 2000 2500 Messages/sec .
Message Loss Ratio 6.00% 5.00% 2-degree Loss Ratio (%) 4.00% 3-degree 3.00% 2.00% 1.00% 0.00% 0 500 1000 1500 2000 Messages/sec
Conclusions and Future Work Initial results indicate overlay networking is a promising strategy for providing many-to-many multicast in the open network environment of DS-RT. We are working on an architecture specification based on the properties of distributed simulation traffic plus recent networking research. NPS is working on a Web-service-based registry and routing information system.
Recommend
More recommend
