1 / 24

GENI System Overview

GENI System Overview. 2008. 10. 20. SeungHo Lee. Outline. Document Scope Introduction Using GENI – An Example GENI System Overview Entities Operations Substrates, Aggregates, and Slices Connecting a Slice to a Non-GENI Network Slice Requests, Authorization, and Audit Information

jenski
Télécharger la présentation

GENI System Overview

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. GENI System Overview 2008. 10. 20. SeungHo Lee

  2. Outline • Document Scope • Introduction • Using GENI – An Example • GENI System Overview • Entities • Operations • Substrates, Aggregates, and Slices • Connecting a Slice to a Non-GENI Network • Slice Requests, Authorization, and Audit Information • GENI Tools & Services • Why Clearinghouses? • GENI Instrumentation and Measurement

  3. Introduction • What is GENI • Global Environment for Network Innovations • A suite of network research infrastructure • Now in its design and prototyping phase • Sponsored by the NSF • Support NetSE experiments since 2005 • Core Concepts for GENI • Programmability • Virtualization and other forms of resource sharing • Federation • Slice-based experimentation

  4. Introduction • Design Goals • Generality • Diversity & Extensibility • Fidelity • Observability • Ease of Use • Sliceability • Controlled Isolaton • Opt-in • Security • Federation & Sustainability

  5. Using GENI – An Example • Resource Discovery Clearinghouse knows 1) what resources exist, 2) which resources are currently available, 3) which are already booked for other researchers, 4) which may be unavailable due to planned and scheduled events, 5) and who is authorized to use resources. Researcher consults a resource discovery portal (GENI clearinghouse).

  6. Using GENI – An Example • Slice Creation • Individual resources must be allocated within an aggregate. • These resources must be woven into a coherent topology within the aggregate. • The aggregates must be stitched together to form a coherent slice.

  7. Using GENI – An Example • Experimentation

  8. Using GENI – An Example • Growing an Experiment (Modifying a Slice) • Some challenges for system design • Resource allocation problem; GENI cannot perform a once-and-for-all global optimization of resource usage across slices. • The means by which an existing computational element “learns” that it has new links. • The technical means employed to manipulate aggregate topologies.

  9. Using GENI – An Example • Federated Facilities

  10. Using GENI – An Example • GENI Operations and Managements

  11. GENI System Overview

  12. Major Entities • Aggregates and Components • Components are the individual “things” that can be obtained and programmed for running experiments. • Aggregates are groups of things owned and administered as an ensemble by some organization. • The component manager provides the interface to the control framework, manages resource allocations, and configures components to provide slivers. • When components are organized into aggregates, an aggregate manager provides the above functions plus any needed organization of components to provide resources that span multiple components. • The aggregate/component manager may apply access control or resource allocation policies. • Most aggregates/components provide an operations portal to export operational data to GENI O&M.

  13. Major Entities • Clearinghouse • The clearinghouse is a software entity that is logically centralized, probably implemented in a distributed fashion for robustness. • The principal registry holds a record for each GENI-associated researcher, PI, administrator, operator, etc. • The slice registry holds a record for each slice, equivalent to a “bank account” for that slice in that it records transactions and authorized accesses. • The component registry holds a record for each affiliated substrate component or aggregate. • A ticket is a “sliver record” that specifies the resources that a component allocates to a given slice.

  14. Major Entities • Research Organizations • GENI includes research organizations (primarily universities) including researchers with associated experiment control tools. • Experiment Support Services • Software storage services • Measurement storage services • Visualization services • Opt-in End Users • GENI Opt-in End Users are those who choose to participate or “opt-in” to a GENI experiment, and become part of the slice. • Including real-world users and traffic in GENI is key to providing the fidelity experimenters need in the GENI infrastructures.

  15. Major Entities • GENI Operations and Management • GENI Operations and Management(O&M) are provided by people, tools, and services who administer and operate some or all parts of the GENI system, and in particular its clearinghouse. • Administration and accounting tools help administrators to provide routing functions. • Operations and management tools help operator to manage the overall GENI system and its interfaces to other systems and to troubleshoot, resolve, and record issues in the suite of GENI infrastructures. • Help desk tools support researchers, PIs, slice administrators in setting up slices, running experiments and reporting and escalating problems.

  16. Operations • Setting up a GENI Suite • A clearinghouse is established. • One or more aggregates register with the clearinghouse. • The registered aggregates publish their resource information to the clearinghouse. • An aggregate can belong to multiple clearinghouses. • So the aggregate is responsible for keeping its clearinghouses up to date. • Research organizations register with the clearing house. • Researchers register with research organizations on the basis of existing or planned experiments. • The clearinghouse itself does not need to authorize individual researchers. • The NetSE Council establishes policy about who may access which resources, and under which constraints. • The clearinghouse federates with other clearinghouses.

  17. Operations • Running an Experiment • The researcher acquires credentials from her research organization. • The researcher sends her credentials to a GENI clearinghouse and requests a slice identifier from the clearinghouse. • The clearinghouse validates that this research organization is indeed trusted and provides the researcher with a globally unique slice identifier. • The researcher queries the clearinghouse for available resources. • The clearinghouse provides her with resource information, including current views and projected schedules. • The researcher contacts each aggregate manager for a resource of interest with detailed queries about available resources. • She presents credentials issued by the clearinghouse that allow her to request resources. • The aggregate manager may apply locally defined policy based on her credentials that will constrain the types and amount of resources the researcher can obtain. • The aggregate manager responds with RSpec describing available resources.

  18. Operations • Running an Experiment (cont’) • The researcher makes a resource reservation contacting the credential issue service in the clearinghouse to acquire a “signed slice credential”. • This is used to get a ticket from the manager of the aggregate/component. • When it is time setup the sliver, the researcher sends the ticket back to the aggregate manager and the resources are made available • The researcher downloads software images into her resources, and starts them running. • The researcher may choose to make changes to the resources used for additional experiment runs. • The slice is torn down and its resources freed. • Researcher request, expiration of the slice lifetime, revocation of a researcher’s credentials, management decision, etc.

  19. Substrates, Aggregates, and Slices • A substrate consists of physical equipment (components) and the management software that manipulates it. • Routers via MPLS connections, or via IP tunnels, or via switched Ethernet, or via direct lightpaths, etc. • An aggregate represents the components in a high-level, researcher-visible form as an ensemble of resources that can be manipulated as a coherent whole. • Multiple sets of CPUs (or virtual machines) and a set of constraints on how links can be formed between these CPUs. • A slice is a “distributed network of programmable elements” (virtual processors, links, etc.) that is instantiated within and across aggregates.

  20. Connecting a Slice to a Non-GENI Network • Reasons why connecting to the Internet • Encouraging users to “opt-in” to an experimental service • Running Internet traffic over an experimental transport system • Running a experimental ISP that peers with real ISPs • Potential Dangers • The experiment will run amok and attack some part of the Internet, whether accidentally or on purpose. • Connection of a slice to the Internet must be undertaken with some caution.

  21. Slice Requests, Authorization, and Audit Information • Resource Specifications • RSpec is a data structure describing a component’s resources • Includes processing capabilities (processor architecture and speed), network interfaces (bandwidth), and privileged operations that can be invoked on the component (access to instrumentation, protected kernel state, and hardware accelerators). • Names & Identifiers • GENI Global Identifiers (GGID) form the basis for a correct and secure system. • GGIDs are used to identify slices, users, and components within the system.

  22. GENI tools & Services • Researcher Tools & Services • Resource Discovery • Experiment Status & Discovery • Debugging • Slice Management • Measurement • Documentation • Code Repository • Operator Tools & Services • Status Monitoring • Identity Management • Failsafe • Resource Allocation • Policy Management • Legacy Internet Services

  23. Why Clearinghouses? • Functions of clearinghouses • Organize Trust Relationships • Between a clearinghouse and research organizations • Between the clearinghouse and aggregate operators • Support Private Versions of GENI • Provide a Framework for Federation • Support Mixed Public / Private Resources for GENI • Provide some portion for use by NSF-sponsored researchers but retain another portion for their own use. • Provide Multiple Resource Allocation Mechanisms • Provide a Graceful Transition Path to ‘Commercial GENI’ • A commercial entity easily set up and operate its own, commercial clearinghouse that charge users and in turns pays aggregate operators.

  24. GENI Instrumentation and Measurement • Requirements for measurement in GENI • Ubiquitous deployment • No (or at least measurable) impact on experiment • Extensibility • High availability • Large capacity • The ability to measure detailed activity with high accuracy and precision from physical layer through application layer • The ability to specify required measurements for an experiment in a slice and then having these measurements initialized in the infrastructure when an experiment is activated. • Access control • A large, secure central repository in which collected data an be anonymized and made available to users • A “data analysis warehouse” where tools for visualizing, interpreting and reporting measurement data can be developed and made openly available

More Related