l -Grids and the OptIPuter Software Architecture

1. l-Grids and the OptIPuter Software Architecture Andrew A. Chien Director, Center for Networked Systems SAIC Chair Professor, Computer Science and Engineering University of California, San Diego Supernetworking Panel SC2003 Phoenix, Arizona November 19, 2003

2. Optical Networking Enables • High Bandwidth • Dedicated Connections • Isolation • Connection Setup And Teardown • So, How Does an Application Use all this Stuff?

3. Exploiting l’s for an Application • Network View: Ad Hoc connections • Applications Request l-Connections • Network Recognizes High BW flows and Configures • System View: Enclave of Resources and Connections • a Distributed Virtual Computer (a SYSTEM) • How to Specify, Implement, and Exploit?

4. OptIPuter Links Three Major Sets ofTechnology Activities • Distributed Virtual Computers • Provide a Simple Abstractions • Aggregate Component Technology Capabilities • Surface Novel Capabilities • High speed Transport Protocols • Long Thread of High Bandwidth-Delay Product Network Protocols • Span The Range “Reach” for Dedicated Optical Connections • Optical Network Signaling and Management • Single Domain and Inter-Domain • Hybrid Circuit and Packet-Switched Networks • Planning and Execution

5. OptIPuter LambdaGrid Software Architecture for Distributed Virtual Computers v1.1 Middleware Protocols Network Config/Mgmt OptIPuter Applications Visualization DVC #1 DVC #2 DVC #3 Higher Level Grid Services Security Models Data Services: DWTP Real-Time Objects Layer 5: SABUL, RBUDP, Fast, GTP Grid and Web Middleware – (Globus/OGSA/WebServices/J2EE) Node Operating Systems Layer 4: XCP l-configuration, Net Management Physical Resources

6. Distributed Virtual Computer Distributed Virtual Computer (DVC) • Distributed Virtual Computer (DVC) • Formed On-demand (use Globus mechanisms) • Dynamic Configuration Of l-network and End Resource Binding • Simplifies Management, Enables Properties • Centralized Resource Control, Security Relations and Operations • Controllable Performance for Distributed Resources (Real-time, QoS)

7. DVC Examples • Virtual Cluster (Hide Complexity of Grid; Resource Flexibility) • Shared Single Domain (Spans Multiple) • Simple Network Naming; Resource Discovery and Access • Private Connections; Simple Performance Characteristics • Real-Time Virtual Cluster for Distributed Collaborative Visualization • Grid Resources + Real-time Network + Real-Time Runtime (TMO) • Collaborative Visualization Cluster • Grid Resources + Unique Displays • Unique Communication Abstractions: Photonic Multicast or LambdaRAM SDSC UCI or UIC SIO/NCMIR UCSD CSE

8. Realizing Distributed Virtual Computers • Research Challenges • Application-driven Definition of Abstractions • Useful Collections which Match Application Paradigms and Needs • Incorporates New Collective Models • DVC Description • Namespaces, Communication, Performance, Real-Time, … • Standard Specifications; Most Applications Parameterize • Integration Of Component Technologies • Executing the DVC on a Grid • Planner That Identifies Resources • Selects from Virtual Grid Resources • Negotiates with Resource Managers and Brokers • Executor and Monitor for DVC • Acquires and Configures • Monitors for Failures and Performance • Adapts and Reconfigures

9. Summary • l’s Novel Capabilities are an Opportunity • High Bandwidth, End to End Pipes, Private Connections • Simpler model than a shared, best-effort network • l’s Novel Capabilities are a challenge • How to manage? How to Fill? • How to Present simply? • OptIPuter Software Research Spans Network Management, Protocols, Middeware, and Visualization • OptIPuter Software Architecture Integrates These Disparate Technologies Into a Simple Model for Applications: “Distributed Virtual Computers” • DVC’s Enable • Transparent use of Optical Network Capabilities • Exploitation of Many of their Advantages