Gigi Karmous-Edwards gigi@mcnc.org May 14, 2006 TERENA Workshop Catania
Outline • Motivation • Optical Control Plane • EnLIGHTened Computing • GLIF • Conclusions
Motivation
E-science and Grid Computing • E-science : global, large scale scientific collaborations enabled through distributed computational and communication infrastructure • Combines scientific instruments and sensors, distributed data archives, computing resources and visualization to solve complex scientific problems • In physics, molecular biology, environmental, Health, Entertainment, etc.
E-science and Grid Computing • Grid computing : main enabler of E-science • Grid is concerned with "coordinated resource sharing and problem solving in dynamic, multi-institutional virtual organizations." (Foster) • E-science migrated to Grid for the reasons of affordability of high-bandwidth communication infrastructure, affordability of resources and inter-disciplinary nature of the research
Demands on Networks: Advanced Support of E-science Apps Grid Cluster Grid Cluster Grid Cluster Grid Cluster Optical Network Grid Grid Cluster Cluster
Advances in Optical Technologies • 1000 channels per fiber….. Experimentation with 160G per channel • Dark Fiber every where …. • Fiber is much cheaper…US Headlines: Google buys Fiber • All-optical switches are getting faster and smaller (ns switch reconfiguration) • Control Plane protocols, SOA, continue to mature • Layer one Optical switches relatively cheaper than other technologies • Electronic Dispersion Compensation • Fiber, optical impairments control, and transceiver technology continue to advance while reducing prices
New Demands on Networks • High bandwidth connectivity of supercomputers (teraflops+) • Large file transfers, over long distances • Advanced support of E-science applications • Application-driven and automatic resource management • Determinism (QoS), jitter and latency requirements • Coordination of network with computational and non- computational resources (CPU, databases, sensors, instruments • Mechanisms for retrieving near-real-time information about network resources and network states • Mechanism for both advance and fast on-the-fly reservation and set-up • Policy and security enforcement in open scientific environments
New Demands on Networks (cont’d) • Applications/end-users/sensors/instruments requesting optical networking resources host-to-host connections - on demand • Near-real-time feedback of network performance measurements to the applications and middleware • Exchange data with sensors via potentially other physical resources • Destination may not be known initially rather only a service is requested from source and the destination is derived from the request information
Research Challenges Integrating Advancing Optical Technologies into the experimental Environment • Advanced reservation of networking resources - Grid Scheduler (middleware) • interacts with control plane Applications requesting optical networking resources – host-to-host connections • (applications interacting w/ control plane (this is not done today) Very dynamic on-demand use of end-to-end networking resources - feedback • loop between control plane, Application, and Grid middleware Near-real-time feedback of network performance measurements to the • applications and middleware Interoperation across Global Grid networks - network interdomain protocols • for Grid infrastructure rather than between operators Policy and Security •
Control Plane
One Definition of Control Plane “Infrastructure and distributed intelligence that controls the establishment and maintenance of connections in the network, including protocols and mechanisms to disseminate this information; and algorithms for engineering an optimal path between end points.” Draft-ggf-ghpn-opticalnets-1
Centralized vs. Distributed… Key areas for Today’s Control Plane are: 1) Provisioning Network 2) Recovery Behavioral Control Network Network Management Management ( Hierarchical ) Migration Protocols Protocols NE NE NE NE NE NE Centralized ( vertical ) Distributed ( Horizontal )
Control Plane Functions • Routing - Intra-domain and Inter-domain 1) automatic topology and resource discovery 2) path computation ( How do we use the infrastructure ) • Signaling - standard communications protocols between network elements for the establishment and maintenance of connections • Neighbor discovery - NE sharing of details of connectivity to all its neighbors ( very powerful tool ) • Local resource management - accounting of local available resources
En LIGHT ened Computing
Highly-dynamic Grid E-science Applications Driving Adaptive Optical Control Plane and Compute Resources NSF seed funded project
En LIGHT ened team significance key Institutions collaborating on the research efforts • MCNC (Network research), -PI -Gigi Karmous-Edwards, Yufeng Xin, Steve Thorpe, Bonnie Hurst, Lina Battestilli, Mark Johnson , John Moore • LSU (Application and Grid research) , PI -Ed Seidel, PI - Gabriele Allen, PI - Seung Jong (Jay) Park , Jon Maclaren, Andrei Hutanu, Lonnie Leger • Renaissance Computing Institute, RENCI (Grid Middleware research): (a joint institute between UNC, Duke and NC State ), PI - Dan Reed, Alan Bletecky, Lavanya Ramakrishnan, Joel Dunn • NCSU (Network research), Savera Tanwir, Harry Perros
En LIGHT ened team significance (cont’d) Key Partner Institutions (cost share) • Cisco , Javad Boroumand, Russ Gyurek, Wane Clark, Kevin McGratten • AT&T Rick Schlichting, John Strand, Matti Hiltunen • IBM Steve Hunter, Ed Bowen • SURA Gary Crane • Naval Research Lab (NRL) , Hank Dardy • Calient Networks , Olivier Jerphagnon, Ron Mackey • UCSB/Calient , John Bowers • NLR
connectivity diagram with partners To Asia To Canada To Europe SEA POR BOI Chicago CAVE wave En LIGHT ened wave (Cisco/NLR) PIT OGD DEN CHI KAN CLE SVL WDC Cisco/UltraLight wave L.A. Raleigh LONI wave TUL San Diego DAL Baton Rouge HOU Official Partners: - AT&T Research Members: NSF Project Partners - SURA International - MCNC GCNS - OptIPuter - NRL Partners - Renaissance Comp. Inst. - UltraLight - Cisco Systems - GLIF - LSU CCT - WAN-in-LAB - Calient Networks - DRAGON - IBM
Participating Applications in several Science areas! 1. Black Hole simulations - Astrophysics LSU 2. SCOOP - Ocean observatory - SURA Partner - Gary Crane 3. BIRN project with Mark Ellisman (NIH) • Optiputer cooperation and EnLIGHTened Wave 4. HEP - UltraLight - Harvey Newman 5. International research - applications with partner NRENs across EU- Enlightened is an official EC project partner (sister-project)
EC Sister project L.U.C.I testbed
Japan’s G-Lambda research collaboration Slide: Courtesy of Michiaki Hayashi KDDI R&D Laboratories Inc.
Japan’s G-Lambda research collaboration Slide: Courtesy of Michiaki Hayashi KDDI R&D Laboratories Inc.
Problem Scope High bandwidth pipes along very long distances – • terabyte transfers, petabyte, etc Dynamic applications adapting to middleware • resource information Network resources coordinated with vital Grid • resources – CPU, and Storage Advanced reservation and on-the-fly dynamic • requests of coordinated resources (CPU,Storage, network)
Problem Scope Deterministic end-to-end connections – low jitter, • low latency Applications requiring both high capacity pipes • and Internet (dual NIC hosts) Near-real-time feedback loop of • Network/CPU/Storage performance measurements and availability to the applications and middleware Global collaboration over global network • resources (GLIF)
Enlightened’s Research Challenges Coordination of resources per request for both on-the-fly and • advanced reservations - Network resources is an integral part of the application’s request for shared resources Advanced reservation in distributed form - Borrow from • ATM research Optimization of Resource Allocation • Interdomain across Global Grid networks - network • interdomain protocols, policies (management plane and control plane, Grid … WEB services ) Dynamic and Adaptive on-demand use of end-to-end • networking resources ( requires near real-time feedback loop )- Identification of functions and interactions between the control plane, management plane, and Grid middleware
Enlightened Research Challenges Monitoring information of resources - i) • identification of information, ii) abstraction of information, and iii) frequency of updates Software algorithms to support multiple classes of • software including highly-dynamic, workflow engines, data-driven and event-driven applications Rethinking the Behavioral Control of Networks • • Control/management planes interacting with middleware • Centralized vs. distributed functionality
Policy Policy Edge Workflow Applications Routers Engines Application Abstraction Layer (API) Translate app request to policy Resource Manager Co-Scheduler Abstraction Feedback Loop Resource Resource •Discovery Monitoring For SLA Allocation •Performance Monitoring •Policy
Middleware Architecture (in -progress)
Conclusions
Recommend
More recommend
