56ca2d60f7cadf4afb7280ac3e346a28.ppt
- Количество слайдов: 27
Global Control of Research Networks Gigi Karmous-Edwards gigi@mcnc. org International task Force panel Fall 2006 I 2 Member Meeting
Outline • GLIF • Global Network Challenges • Control Challenges for GLIF Community • Community of Contributors • Conclusions Fall 2006 I 2 Member Meeting
GLIF www. glif. is Fall 2006 I 2 Member Meeting
Global Lambda Integrated Facility www. glif. is Fall 2006 I 2 Member Meeting
GLIF Control Plane and Grid Middleware Integration wg Mission: To agree on the interfaces and protocols to automate and use the control of the contributed Lambda resources to help users on a global scale access optical resources on-demand or pre-scheduled. Fall 2006 I 2 Member Meeting
GLIF Control Plane and Grid Middleware Integration wg w/ Tech wg 1. 2. 3. 4. 5. 6. Work with GLIF Tech group top establish what are GLIF resources (GOLEs) Defined Network Elements in NDL (network Description Language) Software that reads NDL description Need to write to Google MAP APIs to draw resources on a global bases Provide algorithm to compute path from broker information Establish WEB services for connection services Thanks to Jereon Van Der Ham Fall 2006 I 2 Member Meeting
GLIF Automation? Client A Grid middleware Network Management NREN Control Plane WEB Services ? Network Management NREN Control Plane Client B Grid middleware Network Management NREN Control Plane Fall 2006 I 2 Member Meeting
Global Network Challenges Fall 2006 I 2 Member Meeting
New Global Demands on Networks Emerging High-End applications • High bandwidth connectivity between supercomputers (teraflops+) and Data stores, and Instruments globally • Applications/end-users/sensors/instruments requesting optical networking resources host-tohost connections across multiple networks (NRENS, Research projects) • Coordination of network with computational and non-computational resources (CPU, databases, sensors, instruments ) Fall 2006 I 2 Member Meeting
New Global Demands on Networks Emerging High-End applications (cont’d) • Mechanisms for retrieving near-real-time information about network resources and network states across all global networks • Mechanism for both advance and fast on-thefly reservation and set-up • Policy and security enforcement in open scientific environments Fall 2006 I 2 Member Meeting
Control Challenges for GLIF Community Fall 2006 I 2 Member Meeting
Control Challenges for GLIF Community • Local control vs. Global control • Centralized vs. Distributed control • Diversity in Grid middleware • Interoperation between Grid Projects • 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 Fall 2006 I 2 Member Meeting
Control Challenges for GLIF Community (cont’d) • 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 Fall 2006 I 2 Member Meeting
Applications Edge Routers Workflow Engines Application Abstraction Layer (API) Translate app request to policy Resource Manager Co-Scheduler Policy Abstraction Feedback Loop Resource Allocation Resource Monitoring For SLA Monitoring Fall 2006 I 2 Member Meeting • Discovery • Performance • Policy
Control Plane (Distributed Control) “Carrier view” “Drivers” “E-science view” Applications/ Grid middleware Management Plane Management Control. Plane Operations Support Plane System (OSS) Control. Plane Operations Support System (OSS) Control Plane Networks Data Plane Application Network-wide, global, comprehensive, distributed, fully automated software system that enables interoperability, responsiveness, flexibility, enhanced access to network resources, and, speed and efficiency gain Fall 2006 I 2 Member Meeting
Three Control Entities Client A Grid middleware Network Management NREN Control Plane (NM, CP, Grid Middleware) WEB Services ? Network Management NREN Control Plane Client B Grid middleware Network Management NREN Control Plane Fall 2006 I 2 Member Meeting
Management Plane Functions • Usually rely on client/ server model • Hierarchical management applications communicating to each network element in its domain via a management protocol, (i. e. , SNMP, Tl 1, XML, etc). Centralized Vertical Model Network Management (Hierarchical ) NE NE NE Fall 2006 I 2 Member Meeting
Control Plane Functions • Routing - Intra-domain and Inter-domain, automatic topology and resource discovery, path computation • Signaling - protocols for the establishment and maintenance of connections • Neighbor discovery - NE sharing of details of connectivity to all its neighbors • Local Resource Management - accounting of local available resources NE Protocols NE Distributed Horizontal Model Fall 2006 I 2 Member Meeting
Grid Middleware Functions • Knows the schedules of the network resources • Path Computation/Traffic Engineering • AAA/Policy Enforcement • Reliability and Performance Predictions • Discovery and Resource Registration • Keeps monitoring information about the network Fall 2006 I 2 Member Meeting
Community of Contributors Fall 2006 I 2 Member Meeting
NRENs, Projects and Initiatives from Around the Globe contribute …. . Open to ALL • Please join us and bring your ideas, experience, and issues as we migrate towards linking the world with LIGHT (Kees)! • NRENs • Government Initiatives • Other Standard bodies: OGF, IETF, ITU, etc…. • Individuals and Organizations • Current Research Projects: – Asia - Japan’s G-lambda project, GLORIAD, etc… – North America - Optiputer, Dragon, UCLPv 2, Enlightened, NLR, HOPI, etc…. – EU - Phosphorus, Geant 2, Neather. Light, Star. Plane, etc…. – Others… Fall 2006 I 2 Member Meeting
Rationale behind SC 06 BOF (slides from Dimitra Simeonidou) • A strategic alliance between the two international groups: GLIF control plane and OGF ghpn • Focus: – End-to-end on-demand advanced scheduling of optical network resources for high-end grid applications – Still many outstanding issues: technical, organizational and policy based • Why Now? – A number of funded research projects to contribute experience and effort – Initiatives planning the next generation of research infrastructures – Wide consensus from the community that we need to change the way we design and deploy networks – Standardisation efforts in parallel with OGF Fall 2006 I 2 Member Meeting
Basis for the proposed liaison (slides from Dimitra Simeonidou) • Collect and analyze existing experience on delivering network services across heterogeneous optical domains – Establish strategic relationships with relevant projects and initiatives: • Research projects (i. e. Enlightened-US, PHOSPHORUS-EU, G -Lambda-Jp, Starplane, GEANT 2, UCLPV 2, Optiputer, Viola, e. Photon/ONE+ WP-JP-G, COST 291, NOBEL 2 ) • Initiatives (i. e. EARNEST-EU ) – Critical appraisal of operational and research experience – Technology benchmarking – Deployment roadmaps • Contribute towards the ghpn standardisation effort – Focused documents defining network interfaces among multiple network and Grid layer Fall 2006 I 2 Member Meeting
Conclusions Fall 2006 I 2 Member Meeting
A Book written by the Community Published! Fall 2006 I 2 Member Meeting
Conclusions • Control Plane research is vital to meeting future generation Grid computing - with a strong focus on algorithms to meet the needs of driving applications. • Simplifying Control of Global network resources in coordination with compute relevant resources (clusters, data store, instruments, etc. ) is critical to next generation Scientific and Enterprise advancements. • The Research networks ( w/ exception to Japan’s Carriers) are the ones taking bold steps w/ control plane deployments, testbed infrastructures… Continue to foster collaboration and apply lessons learned to production quickly. • International Collaboration is a very Key ingredient for the future of Scientific discovery - The control of Optical network plays the most critical role in achieving this! Fall 2006 I 2 Member Meeting
Thank You! Gigi Karmous-Edwards gigi@mcnc. org Fall 2006 I 2 Member Meeting
56ca2d60f7cadf4afb7280ac3e346a28.ppt