Transcript Next Generation Internet - network systems lab @ sfu

Next Generation Internet
CMPT 771 – Internet Architecture & Protocols
Presented by: Bassam Almohammadi
Slide 1 of 16
Motivation
What is wrong with the current generation ?
Slide 2 of 16
Old Design
The original design of the Internet
was developed more than 30
years ago.
Before Personal Computers,
Local Area Networks.
•
Limited number of users, mostly
scientists.
•
•
Links speed <= 50 kbps
Different requirements were
considered.
•
Slide 3 of 16
Old Requirements
(and priorities)
1.
Multiplexing
- the primary goal.
- network of sub-networks
2.
Survivability
- robustness
3.
Service Generality
- at transport level
4.
Diverse subnet technologies
- heterogeneity
Slide 4 of 16
Old Requirements
(and priorities)
1.
2.
Multiplexing
- the primary goal.
- network of sub-networks.
5.
Distributed Management.
6.
Cost Effective
Survivability
- robustness.
7.
Ease of Attachment
8.
Accountability.
3.
Service Generality
- at transport level.
4.
Diverse subnet technologies
- heterogeneity .
Slide 5 of 16
Why the design history matters?
“ An understanding of
the history of the design
provides a necessary
context for current design
extensions.”
[David D. Clark, ‘88]
Slide 6 of 16
What has changed?
Requirements – what did we miss?
Slide 7 of 16
What has changed ?
1- The Internet as an economic reality

The full economic implications of the
Internet architecture were not
appreciated in the first design.
“I have traveled the length and breadth of this
country and walked with the best people,
and I can assure you that data processing is a
fad that won't last out the year.”
Editor in charge of business books for
Prentice Hall, 1957
Slide 8 of 16
What has changed ?
1- The Internet as an economic reality

ISPs  1990s.

How the architecture design can influence the industry?


The design of the network architecture induces the structure of the industry

Open protocols among routers → competitive market in routers.

BGP → competitive market in wide-area ISPs.
The next generation cannot succeed without being economically worthy.
Slide 9 of 16
What has changed ?
2- The erosion of trust and the need for security



The size of the network.
Universal transparency is scary!
 The core of the Internet is oblivious to the content of packets.
 It might be an important factor in the success of the Internet.
 However, it becomes prone to security risks.
The size of the network and, the level of transparency and risk.
 Powerful transparent → seriously risky → firewalled !
 Firewalls affected the ease of deploying new applications on the Internet.
Slide 10 of 16
What has changed ?
3- The emergence of conflicting interests
Privacy
V.S.
Lawful interception
Protection from spam and adware
V.S.
Spammers and advertising agencies
Sharing of multimedia
V.S.
Intellectual rights holders

This existence of this tussle might prevent the achievements of essential goals.
(security).

An overarching objective of all architectural decisions : “Design for tussle”.
Slide 11 of 16
What has changed ?
4- New application requirements

Is the “best effort” specification for packet transfer service sufficient enough ?
 The network make no specific commitments about transfer characteristics.
 End-system software must take this unpredictability into account.
 “best effort” made Internet protocols could run “anything”.
 Is this acceptable in a “grown up” Internet ?
 As a commercial product with paying users.
 As a tool for more demanding applications (real-time).

Wide range of operating conditions OR predictability? Tradeoff must be managed.
Slide 12 of 16
What has changed ?
5- New technology features

Since the Internet protocols were first proposed:

High-speed local area networks – new wide-area technologies (ATM) – new
mechanisms for virtualization and encapsulation (MPLS) and wavelength
division multiplexing – the emergence of embedded devices.

It’s remarkable how the Internet has accommodated most of new technologies.

One certain technology that does challenge the original architecture ?
Slide 13 of 16
What has changed ?
5- New technology features

WIRELESS (Mobility)

Some technical features (Higher loss rates) raise specific issues (Congestion Ctrl)

Main issue: Mobility

Raises: addressing, routing, security, service quality variation, etc.
Slide 14 of 16
Architecture then and now
In the light of the changing requirements
Slide 15 of 16
Key architectural features

Packets



Addressing and identity



The basic idea is robust, and has stood the test of time.
Abstraction should be rethought to meet today’s need.
The IP address must be rethought (separating location and identity).
E.g. FARA [FARA03]
Security


Must not be viewed as a single dimensional space (i.e. more is better)
but a multi-dimensional space, shaped by conflicting interests.
Back to the conflict of interests to precisely define security requirements.
Slide 16 of 16
New architectural directions


New Congestion Control scheme

To accommodate new higher speeds.

E.g. XCP (eXplicit Congestion Protocol). [NewArch, 2003]
An alternative to protocol layering

Layered protocol: Increased complexity and rigidity.

Serious architectural limitations (layer violations, feature interaction, proliferation)

What about a non-layered architecture?

E.g. RBA (role-based architecture). [NewArch, 2003]
NGI Projects and Resources









NewArch
European Future Internet
GENI : Global Environment for Network Innovation.
Clean Slate Research Program (Stanford University).
FIRE : Future Internet Research & Experimentation.
Internet 2
Future Internet Forum (Korea).
AKARI (Japan).
G-Lab (Germany).
References

Clark, D., Sollins, K., Wroclawski, J., Katabi, D., Kulik, J., Yang, X., and others,
“New Arch: Future Generation Internet Architecture”. Final Technical Report.
2003.

Clark, D., “The Design Philosophy of the DARPA Internet Protocols”. SIGCOMM
1988.

Clark, D., Sollins, K., Wroclawski, J., Barden, R., “Tussle in Cyberspace: Defining
Tomorrow’s Internet”. ACM SIGCOMM 2002, August 2002.

National Coordination Office for Networking and Information Technology
Research and Development, “NGI: A Glimpse Into the Future”, Indiana
University, April 9, 1998.