FIND

John Wroclawski

USC ISI

IEEE CCW - October 2005

Good Morning

Caveats: Shilling for others: Dave Clark (MIT) and a

collection of collaborators and kibbitzers Not speaking for the NSF 20 minute version of a 60 minute talk.. “We” is you. And me. All of us in the

networking research community

The starting point

NSF is working with its research communities and interested collaborators to create a major new networking research initiative

It has two parts FIND is a research program GENI is a facility for research - a piece of

infrastructure

FIND is the lead motivation for GENI

GENI is broader in use than FIND

This talk is almost entirely about FIND

What needs doing? Help people to “think architecturally” Bring out and develop the best

architectural ideas (of any size..) Coalesce ideas into architectural

proposals Test, evaluate, deploy.. Impact the larger world..

Not quite business as usual.

FIND’s challenge questions…

1) What are the requirements for the global network of 10 or 15 years from now, and what should that network look like?

To conceive the future, it helps to let go of the present:

2) How would we re-conceive tomorrow’s global network today, if we could design it from scratch? This is not change for the sake of change, but a

chance to free our minds. A clean slate process.

Isn’t today’s net good enough?

Security and robustness. As available as the phone system Been trying for 15 years -- try differently?

Easier to manage. Really hard intellectual problem No framework in original design

Recognize the importance of non-technical considerations Consider the economic landscape. Consider the social context.

The technical pushNew network technology

Wireless Mobility Dynamic impairments

Advanced optics Dynamic capacity allocation

New computing paradigms Embedded processors everywhere Sensors.. Grid..

Whatever computing is, that is what the Internet should support. The Internet grew up in a stable “PC paradigm” time. Now it is becoming different..

The scope of the challengeIs it “Internet classic”? A cloud of routers with general

purpose computers at the edges? No! The scope of the question is much bigger than that. Ask: what will “the edge” look like. That is where the action

is. Sensors. Embedded computers

Or is it?Ask: what is it that users do? Try to conceptualize a

network that supports them Information access and dissemination Location management and location-aware systems Identity management systems Conceptualize at a higher level (not higher layer)

What should we reconsider?

For the moment, everything Packets, datagrams, circuits? -- (yawn) Religious beliefs

End to end, transparency, our model for layering, layering…

The F is “Future”. To conceive of a future, must let go of the present This does not mean that we cannot get there

incrementally.. But it is useful to know where you’re going

Defining success

1. We throw away the current Internet. The most dramatic form of success.

2. We set a goal, and the we realize we can get there incrementally.

Impose a bias or direction on change.

3. Lots of fresh ideas leak into the present Internet.

Research community shows up at the IETF again

Timing

This is a long term effort. IPv6 started in 1990.

It is less important when we start, more important that we do so. We can and will do mid-course correction. Adjust the objective as we get closer.

Long term research has short-term fallout.Short term research only accidentally, if at

all, achieves a long-term objective.

A key benefit

Today, we see erosion of clean design principles--architecture. Should we care?

Clean architecture means clean interfaces, as well as better behavior.

Interfaces create opportunities for innovation. Architecture defines a framework around which

innovation and evolution occurs The definition paradox The avoidance of accidental limitations

If we don’t do this?

If we don’t step up to conceive of what networking will be in 10 years: A narrowing of the utility of the Internet to

specific purposes. E-commerce? A pervasive loss of confidence in Internet. Limit ability ability to exploit new technology. A shift of focus (inside NSF) to sectors that

seem more relevant and vigorous. A gentle glide into irrelevance for research.

Caution: Gears Shifting

Architecture

A process: putting components together to make an entity that serves a purpose.

A result: entities come to be defined by their architecture. Think about the original form of architecture.

A discipline: architects study past examples, learn patterns and approaches.

All of these apply to “real architects” and to computer science.

Architecture research areas:

Putting components together: Modularity, interfaces, reuse, dependency

For a purpose: Successful architecture recognizes what a system

cannot do.

Honoring design patterns, approaches and cautions (and increasingly, systems theory). General: layering, abstraction, size of modules,

second-system syndrome. Internet: end to end, transparency vs. conversion

(spanning layer), the “hour-glass model”, soft/hard state.

How does GENI fit in?

Framework for Building Blocks Virtualization, embedder, management

Framework for Services Stable platform, access to users,

measurement and observation

Framework for Transition Real users, connection to current ‘net,

scalable

Framework for Community

GENI: Goals and Key Concepts

Goal: shared platform that promotes innovations

“embedding infrastructure for testbeds”

Key Concepts: Slicing, Virtualization, Programmability, Modularity, Federation..

Details of the Facility (snapshot)

Internet

backbone wavelength

backbone switch

Sensor Network

Edge Site

Wireless Subnet

Customizable Router

Global and Local Software

CM

Node substrate

CM

Node substrate

Components

CM

Node substrate

Resource Controller Auditing Archive

Slice Manager

RDS CSPS LSMS

Infrastructure services

. . . and others . . .

Core

nodecontrol

sensordata