Title: Routing%20Research%20Group@ietf64%20%20Routing%20in%20a%20Future%20Internet%20Architecture%20
1Routing Research Group_at_ietf64Routing in a
Future Internet Architecture Contributions from
the NewArch ProjectBob Braden, USC/ISI
2The NewArch Research Project
- Funded by DARPA July 2000 June 2003.
- Primary Participants
- MIT Dave Clark, John Wroclawski, Karen Sollins
- ISI Bob Braden, Ted Faber, Aaron Falk
- ICIR (Berkeley) Mark Handley
- Other Noel Chiappa
- Grad students Xiaowai Yang, Dina Katabi,
Joanna Kulik, Venkata Pingali - The Internet Architect 1980 1990
- The primary source of this talk .
3The NewArch Project
- Ambition "Develop and evaluate a strengthened
Internet architecture for the 10-20 year time
frame - Results
- Review of changing requirements
- High-level exploration of some key architectural
issues and ideas - Conference and workshop papers
- (slightly monumental) Final Report
- http//www.isi.edu/newarch/
4Some Ideas that NewArch Considered
- Economics (business models,
tussle, etc.) - Trustworthiness (security and robustness)
- Greater architectural heterogeneity
- Ubiquitous mobility
- Expanded application architecture
- Meta-architecture
- http//www.isi.edu/newarch/
5General NewArch Publications
- Tussle paper
- Clark, Wroclawski, Sollins, Braden, Tussle in
Cyberspace Defining Tomorrows Internet. ACM
SIGCOMM 2002. - Architectural Overview
- Clark, Sollins, Wroclawski, Faber, Addressing
Reality An Architectural Response to Real-World
Demands on the Evolving Internet. ACM SIGCOMM
2003 FDNA Workshop. - Final Technical Report
- Clark et. al., NewArch Future Generation
Internet Architecture. Final Technical Report
for DARPA. - See http//www.isi.edu/newarch/
6 Specific NewArch Research Initiatives
- Global internetwork without a global address
space FARA - Clark et al, Trans on Networking, June 2005
- Routing architecture for user empowerment NIRA
- Wang, Sigcomm 2003
- Explicit Congestion Control XCP
- Katabi et al, Sigcomm 2002
- Non-layered architecture (Role-Based
Architecture) - Braden et al, Hotnets I, 2002
- Regionalization
- Sollins et al, Sigcomm 2003 FDNA Workshop
7Routing in NewArch
- Issues we discussed
- Is routing architecture fundamentally
separable from the overall network architecture? - Hypothesis Chiappa the existing routing
paradigms do not scale need map-based routing
(NIMROD). - How do the goals of economic viability and
trustworthiness impact the routing architecture? - Can user empowerment be achieved with scalable
routing?
8Goal User Empowerment
- Give end users (some) control over
- Inter-domain routing
- Use inter-ISP competition to encourage
introduction of improved network-level services
(e.g., QoS, multicasting, ) - Application-layer elements within the network.
- Relays, proxies, firewalls, other middle boxes.
9Routing for User Empowerment
- Xiaowei Yang's MIT thesis described an
architecture for user-empowered inter-domain
routing NIRA. - NIRA A New Internet Routing Architecture,
Xiaowei Yang, ACM Sigcomm 2003 FDNA Workshop,
August 2003. - NIRA represents a significant rethinking about
inter-domain routing. - It is RESEARCH probably not final answer, but
important ideas.
10Acknowledgment
- The rest of this presentation is heavily based
upon Xiaoweis slides, with her kind permission. - Xiaowei is now on the CS faculty of UC Irvine.
- The RRG might invite Xiaowei to present her own
work.
11We Want to Let Users Choose Domain-Level Routes
ATT
UUNET
- Our hypothesis
- User choice stimulates competition.
- Competition fosters innovation.
- Validation requires market deployment.
- NIRA the technical foundation.
Local ISP
12Central Ideas of NIRA
- Built on earlier ideas of explicit routing,
up/down routing. - Defines efficient representation of explicit
route for common case. - Assuming today's generally tree-shaped
inter-domain topology, with providers and
customers - "Core" in the center.
- Strict provider-rooted hierarchical addressing
13NIRAs Addressing
Core
B2
B1
1/16
2/16
13/32 21/32
- Strict provider-rooted hierarchical addressing
- Using fixed-length sub-field of prefix per level
- An address represents a valid route to the core.
11/32
12/32
R3
R2
R1
131/48 211/48
111/48 121/48
N3
122/48
N1
N2
1111000 1211000
1312000 2112000
14NIRA Addressing
- A source and a destination address unambiguously
represent a common type of route (strictly up
down) - General routes may use source routing headers
- Scalable because
- Financial factors limit the size of core.
- Provider hierarchy is shallow.
- A domain has a limited number of providers.
- Leaving out a number of important details see
paper.
15Basic Forwarding Algorithm
Core
B2
B1
1/16
2/16
13/32 21/32
11/32
- Uphill and downhill routing tables in each
router. - Look up destination address in the downhill table
(for short-cut). - If no match, look up the source address in the
uphill table.
12/32
R1
R3
R2
131/48 211/48
111/48 121/48
N3
122/48
N1
N2
1312000 2112000
1111000 1211000
16 System Components of NIRA
- Addressing
- Route discovery
- Topology Information Propagation Protocol (TIPP)
- A user learns his addresses and topology
information (static) and perhaps route
availability (dynamic) - Name-to-Route mapping
- Name-to-Route Lookup Service (NRLS) an enhanced
DNS service - A user learns destinations addresses and
optional topology information. - Combining information from TIPP and NRLS, a user
is able to select an initial route.
17What does TIPP Do?
Core
B1
B2
X
- Propagates addresses and topology (controlled by
policy) - May proactively notify of route unavailability
- User can then construct partial topology map of
HIS routes to core.
R1
R3
R2
N3
N1
N2
1111000 1211000
temporarily unusable
18Topology Map -- Bob's View
Core
B1
- User can then construct partial topology map of
HIS routes to core.
R1
R3
R2
N1
1111000 1211000
19Name-to-Route Lookup Service (NRLS)
Foo.com server
Core
Alice.foo.com 1312000 2112000
B2
B1
R1
R3
R2
N3
N1
N2
1111000 1211000
Alice.foo.com 1312000 2112000
20NIRA Summary
- User choice
- Choosing addresses ? choosing routes ? choosing
providers - Failure Handling
- Reactive Unreachable or timeout gt new route
- Proactive TIPP may signal route failure
- No global routing knowledge
- Consideration of provider compensation mechanisms
- "Evaluation shows NIRA is feasible."
21Issues for the RRG
- Is user empowerment a good idea, in general?
- Should user empowerment be a requirement for
inter-domain routing, in particular? - Is the NIRA approach a fruitful direction?
- NIRA A New Internet Routing Architecture,
Xiaowei Yang, ACM Sigcomm 2003 FDNA Workshop,
August 2003.
22 23NewArch Final Technical ReportConclusions
- Future architecture should design for
- User Empowerment
- Tussle SIGCOMM 2002
- Trust-modulated transparency
- Packet aggregates as a network abstraction
- Varying amounts of control state in the network
(not only datagrams) - Freedom from requiring a global address space
- Application control over network services QoS,
etc. - Ease of user configuration
24NewArch Conclusions
- Retain fundamental Internet nature
- An application-independent data carriage service
based on variable-length packets. - But, re-examine some assumptions
- Pure datagrams
- Global addresssing
- IPs linkage of location with identity
- Universal transparency
- Network-layer multicast is fundamental
25What is Network Architecture?
- Network architecture defines FTR 2.
- What the network is for.
- How it accomplished those functions.
- How the system is broken into fundamental objects
- How those parts interact assumptions each part
makes. - Also, it should be explicit about what is NOT
specified. - Meta-architecture
- FARA defines a class of architectures.
Instantiated one specific FARA-based
architecture, FARA-M FARA paper 2003 FDNA
26Routers Forwarding Tables
Uphill table
Core
B2
B1
1/16
2/16
1/16 B1
- Uphill table providers
- Downhill table customers, self
- Bridge table all others
13/32 21/32
11/32
12/32
Downhill table
R1
R3
R2
111/48 121/48
131/48 211/48
111/48 N1
11/96 self
N3
122/48
N1
N2
1111000 1211000
1312000 2112000
27Provider Compensation
- Contractual agreements.
- Users cannot use arbitrary routes.
- Providers use policy checking to prevent
illegitimate route usage. - Direct business relationships
- Common case verify source address
- General case packet filtering
- Indirect business relationship end users pay
remote providers - Policy is made upon the originator or the
consumer of a packet. - An open and general problem. Working in progress
with Jennifer Mulligan and David Clark. - Various billing schemes are possible.
- Flat fee, usage-based billing.
28Related Work
- Routing architecture proposals
- Nimrod RFC 1992, Inter-Domain Policy Routing
RFC 1478, Scalable Inter-Domain Routing
Architecture Estrin92, TRIAD Cheriton00,
Feedback Based Routing Zhu02 - Addressing schemes
- PIP Francis93, SIP Deering93, IPv6 RFC
3513, Provider-rooted addresses Tsuchiya91
Francis94 - My work
- Supports user choice
- Scalable
- Evaluation
29Looking Forward
- New provider compensation model
- Packet-level authentication
- Stable routing with user choice
- Deployment of NIRA
- Tools, methodologies, and principles to guide
system design and understanding - Benchmark simulation configurations
- Robust protocol design