38
Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009 www.geni.net

Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Embed Size (px)

Citation preview

Page 1: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation

GENIStatus and Outlook

Chip ElliottJuly 21, 2009

www.geni.net

Page 2: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 2July 21, 2009

Outline

• Current status• GENI Solicitation 2• “Meso-scale” prototyping• Spiral 2

Page 3: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 3July 21, 2009

Chip Elliott (GPO)

NetSE Council

Ellen Zegura (Chair) Tom Anderson (UW) Joe Berthold (Ciena) Charlie Catlett (Argonne) Mike Dahlin (UT Austin)

Joan Feigenbarum (Yale) Stephanie Forrest (UNM) Jim Hendler (RPI) Michael Kearns (U.Penn) Ed Lazowska (UW) Peter Lee (CMU)

Larry Peterson (Princeton) Jennifer Rexford (Princeton) Alfred Spector (Google)

And not shown . . .

Roscoe GilesHelen Nissenbaum

Page 4: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 4July 21, 2009

GENI Conceptual DesignInfrastructure to support at-scale experimentation

Mobile Wireless Network Edge Site

Sensor Network

Federated International Infrastructure

Programmable & federated, with end-to-end virtualized “slices”

Heterogeneous,and evolving over time viaspiral development

Deeply programmableVirtualized

Page 5: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 5July 21, 2009

Spiral DevelopmentGENI grows through a well-structured, adaptive process

GENI Prototyping Plan

Use

Planning

Design

Build outIntegration

Use

• Spiral Development ProcessRe-evaluate goals and technologies yearly by a systematic process, decide what to prototype and build next.

• Each spiral is 12 months long

• Spiral 1 . . .October 1, 2008 - October 1, 2009

we are here

• Spiral 2 . . .October 1, 2009 - October 1, 2010

Page 6: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 6July 21, 2009

FederationGENI grows by “gluing together” heterogeneous infrastructure

Goals: avoid technology “lock in,” add new technologies as they mature, and potentially grow quickly by incorporating existing infrastructure into the overall “GENI ecosystem”

NSF parts of GENI

Backbone #1

Backbone #2

Wireless#1

Wireless#2

Access#1

CorporateGENI suites

Other-NationProjects

Other-NationProjects

ComputeCluster

#2

ComputeCluster

#1

My experiment runs acrossthe evolving GENI federation.

My GENI Slice

This approach looks remarkably familiar . . .

Page 7: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 7July 21, 2009

Infrastructure examples in Spiral 1

DRAGON core nodesMid-Atlantic Crossroads WAIL, U. Wisconsin-Madison DieselNet, U. Mass Amherst

ViSE, U. Mass Amherst SPPs, Wash U. ORBIT, Rutgers WINLAB

Page 8: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 8July 21, 2009

World-class expertise in GENI Partners

Internet2 and National Lambda Rail

40 Gbps capacity for GENI prototyping on two national footprintsto provide Layer 2 Ethernet VLANs as slices (IP or non-IP)

National Lambda RailUp to 30 Gbps nondedicated bandwidth

Internet210 Gbps dedicated bandwidth

Page 9: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 9July 21, 2009

Key Goals for GENI Spiral 1Drive down critical technical risks in GENI’s concept

GENIClearinghouse

Components

Aggregate AComputer Cluster

Components

Aggregate BBackbone Net

Components

Aggregate CMetro Wireless

Create my slice

Goal #1Fund multiple, competing teams to develop GENI Clearinghouse technology, encourage strong competition within the first few spirals

Goal #1Fund multiple, competing teams to develop GENI Clearinghouse technology, encourage strong competition within the first few spirals

Goal #2Demonstrate end-to-end slices across representative samples of the major substrates / technologies envisioned in GENI

Goal #2Demonstrate end-to-end slices across representative samples of the major substrates / technologies envisioned in GENI

Page 10: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 10July 21, 2009

Key goals are being metRemarkably fast progress in Spiral 1

• Several clusters are now demonstrating . . .– Control frameworks that manage multiple types of infrastructure– End-to-end slices across a wide range of technologies

• Annual project reviews well underway– Sitting down together to ensure shared understanding of

milestones, achievements, challenges– Planning “summer push” to achieve Spiral 1 goals– Starting to plan Spiral 2– Almost all projects have now been reviewed– GPO hopes to exercise options without funding gaps

• Weds, 12:15 – GPO update on reviews & Year 2 funding

Page 11: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 11July 21, 2009

GENI Spiral 1

• Provides the very first, national-scale prototype of an interoperable infrastructure suite for Network Science and Engineering experiments

• Creates an end-to-end GENI prototype in 6-12 months with broad academic and industrial participation, while encouraging strong competition in the design and implementation of GENI’s control framework and clearinghouse

• Includes multiple national backbones and regional optical networks, campuses, compute and storage clusters, metropolitan wireless and sensor networks, instrumentation and measurement, and user opt-in

• Because the GENI control framework software presents very high technical and programmatic risk, the GPO has funded multiple, competing teams to integrate and demonstrate competing versions of the control software in Spiral 1

Nothing like GENI has ever existed; the integrated, end-to-end, virtualized,and sliceable infrastructure suite created in Spiral 1 will be entirely novel.

Page 12: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 12July 21, 2009

Outline

• Current status• GENI Solicitation 2• “Meso-scale” prototyping• Spiral 2

This section describes a GPO proposal to NSF which is currently under review, and which may or may not be funded.

This section describes a GPO proposal to NSF which is currently under review, and which may or may not be funded.

This is very important

Page 13: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 13July 21, 2009

Goals of Solicitation 2

Filling Gaps

Security GENI’s security requirements and architecture have not yet been formulated with sufficient scope and clarity to enable creation of robust system and subsystem engineering plans.

Experiment Workflow Researchers will need easy-to-use, comprehensive tools to create, debug, and monitor experiments using the GENI infrastructure suite. Good “experiment workflow” tools and systems will be critical to GENI’s success.

Instrumentation and Measurement

Researchers will need good mechanisms for instrumenting their GENI experiments, and archiving, analyzing, and sharing the resultant measurements. Measurements of interest range from CPU and memory utilization, through network packet counts and ambient RF spectrum measurements, to recording arbitrary streams of data such as logging information, state machine transitions, etc.

Building upon Spiral 1

Early Experiments “Shakedown” experiments using GENI prototype systems must begin within 6 months, to provide initial feedback on their utility and help guide design going forward.

Federation of Clearinghouses

Federation is fundamental to the GPO’s plans for creating a GENI “ecosystem” that is larger than NSF-funded portions of an interoperable infrastructure suite.

Augmentation of System Integration Efforts

Cluster leaders will require more labor funding as additional research teams integrate into their clusters, i.e., for additional effort not already funded by Solicitation 1.

Page 14: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 14July 21, 2009

Proposals received in Solicitation 2

Solicitation 2 Proposals

0 2 4 6 8 10 12

Security (large)

Experiment workflow (large)

Inst & Measurement (large)

Early experiments (large)

Aggregages - wired (large)

Aggregates - wireless (large)

Integration / Augmentation

Security (small)

Exp workflow (small)

Inst & Meas (small)

Early experiments (small)

Aggregates (small)

Control framework (small)

Federation (small)

Education (small)

Participation (small)

# proposals receivedTotal: 91 proposals received @ $48M

InstrumentationExperiments

How was the coverage?

WorkflowSecurity

Page 15: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 15July 21, 2009

GPO Selections

Category Number of Large Proposal Number of Small Proposal

Security 5 3

Instrumentation and measurement 5 1

Experiment workflow 4

Early experiments 2

Augmented integration and federation 5 1

Additional aggregates 3 2

Miscellaneous 2

Provide excellent coverage in filling GENI’s critical gaps

Jump-start trials of international federation with leading research teams in Europe, Japan, South Korea, and Brazil

Substantially augment GENI with significant cloud computing capabilities at a very low cost, including OpenCirrus cloud services (currently 768 cores, 10 TB storage) in the PlanetLab cluster, and Amazon’s EC2, S3, and EBS services in the ORCA cluster.

Mesh smoothly into Spiral 1 clusters

This slide describes a GPO proposal to NSF which is currently under review, and which may or may not be funded.

Page 16: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 16July 21, 2009

MetaOps

GENISecurity

Cluster A: TIED

TIED

SecurityProgram

[1720]

TIED

Current Solicitation 1 Projects Added Solicitation 2 Projects

Control Framework

Experiments andWorkflow

Instrumentation

Aggregates

Ops andMgmt

TIED

Legend

Security

Servers

ControlFramework Federation

Switch NetworkGateway WirelessNode

ExperimentWorkflow Experiments Measurement

CloudComputing

OperationsThis slide describes a GPO proposal to NSF which is currently under review, and which may or may not be funded.

Page 17: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 17July 21, 2009

AugmentedFederation

[1756]

Overlay Hosting

PlanetLab

Enterprise GENI

GUSH

MetaOps

Mid-Atlantic

GpENI

ProvisioningService

GENISecurity

Cluster B: Planet Lab

SecurityProgram

[1720]

PlanetLab SecureUpdate[1696]

netKarma[1706]

Current Solicitation 1 Projects Added Solicitation 2 Projects

VM Introspection[1706]

Open Cirrus[1779]

SCAFFOLD[1759]

Control Framework

Experiments andWorkflow

Instrumentation

Aggregates

Ops andMgmt

PlanetLab

Legend

Security

Servers

ControlFramework Federation

Switch NetworkGateway WirelessNode

ExperimentWorkflow Experiments Measurement

CloudComputing

Operations

This slide describes a GPO proposal to NSF which is currently under review, and which may or may not be funded.

Page 18: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 18July 21, 2009

ProgramEdge Node

MetaOps

GENISecurity

IdentityAuthorization

[1785]

Cluster C: ProtoGENI

D Tunnels

CMU Lab

SecurityProgram

[1720]

ProtoGENIAugment

[1743]

Current Solicitation 1 Projects Added Solicitation 2 Projects

EmulabTools[1780]

On TimeMeasure

[1764]

InstrumentTools

MeasureSystem

ProtoGENI

PrimoGENI[1766]

HiveMind

[1792]

MonitoringClusters

[1723]

LeveragingperfSONAR

[1788]

CRON[1794]

ProtoGENI

WirelessExperiments

[1791]

DavisSocial Links

[1780]

Control Framework

Experiments andWorkflow

Instrumentation

Aggregates

Ops andMgmt

ProtoGENI

Legend

Security

Servers

ControlFramework Federation

Switch NetworkGateway WirelessNode

ExperimentWorkflow Experiments Measurement

CloudComputing

Operations

This slide describes a GPO proposal to NSF which is currently under review, and which may or may not be funded.

Page 19: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 19July 21, 2009

ORCAExtend[1700]

GENISecurity

Cluster D: ORCA

DOME ViSE KANSEI

ORCAExtend[1700]

SecurityProgram

[1720]

ORCA

MetaOps

Integ MeasureFramework

[1718]

MeasureLEARN[1733]

BENiGENI[1719]

Current Solicitation 1 Projects Added Solicitation 2 Projects

Cloud Control[1709]

Control Framework

Experiments andWorkflow

Instrumentation

Aggregates

Ops andMgmt

Legend

Security

Servers

ControlFramework Federation

Switch NetworkGateway WirelessNode

ExperimentWorkflow Experiments Measurement

CloudComputing

OperationsThis slide describes a GPO proposal to NSF which is currently under review, and which may or may not be funded.

Page 20: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 20July 21, 2009

GENISecurity

Cluster E: ORBIT

ORBIT ProgrammableRadio[1803]

WiMAX

SecurityProgram

[1720]

ORBIT

Current Solicitation 1 Projects Added Solicitation 2 Projects

ORBIT

MetaOps

Control Framework

Experiments andWorkflow

Instrumentation

Aggregates

Ops andMgmt

Legend

Security

Servers

ControlFramework Federation

Switch NetworkGateway WirelessNode

ExperimentWorkflow Experiments Measurement

CloudComputing

OperationsThis slide describes a GPO proposal to NSF which is currently under review, and which may or may not be funded.

Page 21: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 21July 21, 2009

Emerging Spiral 2 GENI footprint

This slide describes a GPO proposal to NSF which is currently under review, and which may or may not be funded.

Page 22: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 22July 21, 2009

How do the new projects mesh in?

• When do the new contracts start?– Reminder: GPO proposal to NSF is under review;

by no means a sure thing; range of possible outcomes– Our estimate – September (2009)– If funded, we will negotiate with new projects, then fund– IPR agreements were slow in Spiral 1, but we may have sped it up

by requiring letter from contracts office in proposals

• How does integration work?– New projects will begin integration during Spiral 2– Should not have huge effect on most current projects– Except for control frameworks– They already know about the potential work, and their funding will

increase to help with significant increase in workload

Page 23: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 23July 21, 2009

Outline

• Current status• GENI Solicitation 2• “Meso-scale” prototyping• Spiral 2

This section describes a GPO proposal to NSF which is currently under review, and which may or may not be funded.

This section describes a GPO proposal to NSF which is currently under review, and which may or may not be funded.

This is very important

Page 24: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 24July 21, 2009

Meso-scale prototyping

Concept and / orexisting equipment

“Fast track”prototype systemson commercial hardware

“Meso-scale” GENIprototype infrastructure builtwith commercial hardware

“RealGENI”

Spiral Development

and experimentation

This proposal

GENI prototyping

Rapid progress in GENI prototyping has created a remarkable opportunity to accelerate the creation of an end-to-end GENI infrastructure suite for “meso-scale” experiments, leveraging GENI-enabled commercial hardware, across more than a dozen campuses and two national research backbones

This slide describes a GPO proposal to NSF which is currently under review, and which may or may not be funded.

Page 25: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 25July 21, 2009

Benefits of meso-scale prototyping

• Create a compelling infrastructure for entirely new forms of network science and engineering experimentation at a much larger scale than has previously been available

• Stimulate broad community participation and “opt in” by early users across 13 major campuses, which can then grow by a further 21 campuses as the build-out progresses, with a strong partnership between researchers and campus infrastructure operators

• Forge a strong academic / industrial base by GENI-enabling commercial equipment from Arista, Cisco, HP, Juniper, and NEC, with software from AT&T Labs and Nicira.

This slide describes a GPO proposal to NSF which is currently under review, and which may or may not be funded.

Page 26: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 26July 21, 2009

More than a dozen leading US universities

Meso-scale GENI campus prototypes

RutgersOpenFlow & WiMax

PolytechWiMax

WisconsinOpenFlow &

WiMax

ColumbiaWiMax

UMassWiMax

UCLAWiMax

UC BoulderWiMax

U. Wash.OpenFlow

PrincetonOpenFlow

IndianaOpenFlow

GA TechOpenFlow

ClemsonOpenFlow

StanfordOpenFlow &

WiMax

This slide describes a GPO proposal to NSF which is currently under review, and which may or may not be funded.

Page 27: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 27July 21, 2009

WINLAB leadWiMax campus prototypes

RutgersOpenFlow & WiMax

PolytechWiMax

WisconsinOpenFlow &

WiMax

ColumbiaWiMax

UMassWiMax

UCLAWiMax

UC BoulderWiMax

StanfordOpenFlow &

WiMax

NEC WiMaxBase Stations

This slide describes a GPO proposal to NSF which is currently under review, and which may or may not be funded.

Page 28: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 28July 21, 2009

Stanford leadOpenFlow campus prototypes

RutgersOpenFlow & WiMax

WisconsinOpenFlow &

WiMax

U. Wash.OpenFlow

PrincetonOpenFlow

IndianaOpenFlow

GA TechOpenFlow

ClemsonOpenFlow

StanfordOpenFlow &

WiMax

This slide describes a GPO proposal to NSF which is currently under review, and which may or may not be funded.

Page 29: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 29July 21, 2009

OpenFlow backbone prototypesthrough Internet2 and NLR (notional)

ChicagoNLR

AtlantaInternet2, NLR

Salt Lake CityInternet2

DCInternet2

Kansas CityInternet2

HoustonInternet2

SunnyvaleNLR

SeattleNLR

DenverNLR

This slide describes a GPO proposal to NSF which is currently under review, and which may or may not be funded.

Page 30: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 30July 21, 2009

OpenFlow prototypes(current plans,1 of 2)

This slide describes a GPO proposal to NSF which is currently under review, and which may or may not be funded.

Arista 7124S Switch

Rutgers

NLR

Very appealing hardware platform (cheap, fast)

Strong links to Silicon Valley venture capital community

Cisco 6509 Switch

Clemson

Rutgers

Mild endorsement and commitment from Cisco in proposal

HP ProCurve 5400 Switch

Stanford

Georgia Tech

Indiana University

Princeton (eval)

UW Madison (eval)

U. Washington

Internet2

Stanford has demonstrated OpenFlow on HP ProCurve switches

Strong endorsement from vendor in proposal

ProtoGENI is deploying HP ProCurve switches in Internet2 backbone (3 sites this summer)

Page 31: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 31July 21, 2009

OpenFlow prototypes(current plans, 2 of 2)

This slide describes a GPO proposal to NSF which is currently under review, and which may or may not be funded.

Juniper MX240 Ethernet

Services Router

Clemson

Strong endorsement from vendor in proposal

Note vendor commonality with ShadowNet proposal (in Internet2 backbone)

NEC IP8800 Ethernet Switch

Georgia Tech

Princeton (eval)

Rutgers

UW Madison (eval)

Stanford has demonstrated OpenFlow on NEC switches

Strong endorsement from vendor in proposal

NEC also supplying OpenFlow-enabled WiMax basestations for campus builds

Page 32: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 32July 21, 2009

ShadowNet prototypeInternet2 backbone (ProtoGENI sites)

Kansas CityShadowNet

AtlantaShadowNet

DCShadowNet

Salt Lake CityShadowNet

Juniper M7i Routers

This slide describes a GPO proposal to NSF which is currently under review, and which may or may not be funded.

Page 33: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 33July 21, 2009

Outline

• Current status• GENI Solicitation 2• “Meso-scale” prototyping• Spiral 2 - October 1, 2009 to October 1, 2010

Page 34: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 34July 21, 2009

Spiral 2’s central goal

• By end of Spiral 2, GENI control frameworks should be running a significant number of real experiments

• Many major implications for the prototypes– Fairly continuous operation (90% availability?) is hard– Security architecture and mechanisms must be in place & in use– And we need . . .

• Reasonably friendly user interface for researchers

• We need documentation, and probably training tutorials

• Sample experiments

• Very hard, but we must do it

Page 35: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 35July 21, 2009

Ongoing efforts

• Continue integration of control frameworksand aggregates

• Increase interoperability(avoid balkanization of control frameworks)

• Build up instrumentation & measurement tools

• Address Identity Management for researchers(GPO believes Shibboleth is a good starting point, but community must come to a conclusion)

Page 36: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 36July 21, 2009

So what does this mean – for you?

• GPO is starting detailed Spiral 2 milestone discussions– For clusters as groups, and for each individual project– As for Solicitation 2 projects . . . stay tuned . . .

• If you don’t understand everything, don’t panic– You don’t have to solve all the GENI issues– Decide what your project is good at, emphasize that– Keep doing your planned work– Don’t expand to fill the universe

• If you are just joining the GENI project– The GPO will work closely with you to help out– We understand that there is a learning curve

• Bottom line: other people should be using your stuff

Page 37: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 37July 21, 2009

Get involved!

Get your campus involved Talk to GPO

Get your own testbed integrated into GENI

Talk to project PIsand/or GPO

Get your company or organization involved

Talk to project PIs

and/or GPO

International collaborations GPO will be glad to help make connections

Page 38: Sponsored by the National Science Foundation GENI Status and Outlook Chip Elliott July 21, 2009

Sponsored by the National Science Foundation 38July 21, 2009

GENI is now truly underway

• Spiral 1 looks like a success– Several clusters are already demonstrating key goals

– Not every project will succeed . . .

– . . . but a large majority are doing very well indeed

• Spiral 2 will be challenging but a huge opportunity– GPO believes the individual projects are ambitious but achievable

– Integration is likely to have some successes, some failures

– The critical goal is to get real research experiments running

• GENI is now well and truly underway !