Overview and Status Update of

1
Overview and Status Update of
Hui ZhangCarnegie Mellon
2
Slides Used In First Retreat

• Vision
• 100 Mbps to 100 million households
• 1Gbps to 1 million business
• Approach clean slate, design from first
  principles
• Holistic and inter-disciplinary design
• Consider the network as a whole
• Access, metro, core regions
• Service and physical transport layers
• Consider technology trends for scaling, cost,
  future-safeness
• Archiect with explicit considerations of
  economic, dependability, security, mult-service
• Design with explict goals of enabling tractable
  analysis and modeling

3
What is the 100x100 Project?

• Elevator Speech
• What is 100x100 project up to?
• What is the meat?
• What are the key technical challenges?
• What are the key insights and expected
  contributions?
• Is 100x100 project just about speed?
• Why 100 Mbps?
• It is too slow, it is too fast
• Why cannot IP do 100x100?

4
Clean Slate Design

• Why is Clean State Design necessary?
• Have you established that incremental evolution
  is impossible?
• Is Clean State Design useful?
• Any hope of introducing a revolutionary
  solution?
• Mainstream views
• incremental fixing of Internet is sufficient
• radical new solutions have little chance of being
  deployed

5
Broadband AccessOpportunity of An Century

• Vision
• 100 Mbps to 100 million homes
• 1Gbps to 1 million small business
• Fiber and wireless are the access technologies
  of choice of the future

6
Context

• IP is a great success because its architects set
  the right goal global best-effort reachability
• Global addressing scheme
• Internetworking architecture
• Simple best-effort service
• Success is a double-edged sword ? the world
  demands more from IP and the Internet
• Convergence telecommunication infrastructure that
  provides 7x24x365 service
• Dependability, scale, security, economic
  sustainability

7
Convergence Vision
VoIP Data Video over IP
Long
-
haul DWDM backbone
-
POP
(OC48/192)
8
Have We Already Achieved Convergence?
Long Haul
Metro Core
Subscriber/ Enterprise
Metro Access
Service Node/ASP
ISP
Voice Switch
Server
Backbone Router
Metro Hub Office
End Office/ Collocation
Router
Voice Switch
Server
Router
Services
ATM
Transport
RF Cable Copper Fiber
OXC
ACCESS
INTEROFFICE
INTERCITY
G(SONET)
G(?)
Wireless
HAN
9
IP Robustness?
UUNet 10/03/02 Outage as seen from ATT
State of the Art is lacking in robustness on
every time scale days, hours, minutes, seconds
10
Molasses

• Data Plane
• Distributed routers forwarding packets
• Based on FIB or labels

11
A Study of Operational Production Networks(Joint
Between ATT and CMU)

• Obtained anonymized configuration files for 31
  active networks (gt8,000 configuration files)
• 6 Tier-1 and Tier-2 Internet backbone networks
• 25 enterprise networks
• Sizes between 10 and 1,200 routers

12
Configuration State for One Network
13
Example Router Configuration File
14
Complex Interaction of States
Management Plane
Control Plane
Data Plane
Packet Filters
Logic to Combine OSPF and BGP RIBs
Legends
FIBs
Hardwired State
Configuration State
Dynamic State
State Dependency
15
Molasses

• Management Plane
• Figure out what is happening in network
• Decide how to change it

Shell scripts
Tomography
Planning tools
Databases

• Control Plane
• Multiple routing processes on each router
• Each router with different configuration program
• Huge number of control knobs metrics, ACLs,
  policy

• State everywhere!
• Dynamic state in FIBs
• Configured state in settings, policies, packet
  filters
• Programmed state in magic constants, timers
• Many dependencies between bits of state
• State updated in uncoordinated, decentralized way!

Packet filters
Link metrics

• Data Plane
• Distributed routers forwarding packets
• Based on FIB or labels

16
Systems of Systems

• Systems are designed as components to be used in
  larger systems in different contexts, for
  different purposes, interacting with different
  components
• Example OSPF and BGP are complex systems in its
  own right, they are components in a routing
  system of a network, interacting with each other
  and packet filters, interacting with management
  tools
• Complex configuration to enable flexibility
• The glue has tremendous impact on network
  performance
• State of art multiple interactive distributed
  programs written in assembly language
• Lack of intellectual framework to understand
  global behavior

17
Development of the Elevator Pitch

• Best-effort service model is a critical reason
  for Internets success
• Success is a double-edged sword
• Expectation with true two-way broadband
  connectivity to every household (100x100), IP be
  convergence telecommunication infrastructure that
  provides 7x24x365 service
• Far from being dependable, scaleable, secure,
  economic sustainable
• Even when providing best-effort service, it is
  already extremely complex

18
What is the Solution?

• Simplify,
• Simplify,
• simplify!
• But based on what?

19
Observations that Lead to Simplifications

• Tremendous focus on protocols/protocol
  architecture, should also focus on
• networks and network architecture
• Interplay between network architecture and
  protocol architecture
• Architecture agnostic to technology trend
• Should leverage technology trends that enable
  simple network architecture
• Box/Switch/Router centric distributed control has
  fundamental limitations, should also explore
• Control architecture that can implement a wide
  range of network-wide policies

20
Protocol Architecture vs. Network Architecture

• Protocols TCP, IP, BGP, OSPF
• They work in arbitrary network, but do not work
  as well in any network
• Dependable network requires
• Structured network, and
• protocols that can take advantage of network
  architecture

21
Structured Access/Metro Networks
22
Structured Backbone Networks
23
Structured Network Interconnection
Backbone Network
Access Routers
CPE Routers
24
Observation One

• Structured network protocols taking advantage
  of network structures achieve
• Dependable, simple, and understandable network
• Dependable, simple, and understandable protocols

25
Observations that Lead to Simplifications

• Tremendous focus on protocols/protocol
  architecture, should also focus on
• networks and network architecture
• Interplay between network architecture and
  protocol architecture
• Architecture agnostic to technology trend
• Should leverage technology trends that enable
  simple network architecture
• Box/Switch/Router centric distributed control has
  fundamental limitations, should also explore
• Control architecture that can implement a wide
  range of network-wide policies

26
Key Technology Trends

• Trend 1 Deep fiber deployment key to achieving
  ubiquitous, high capacity connectivity to home
• Wireless important complementary technology
• Copper cable can also be used for last 1000
  feet
• Fiber enables not only scalable high speed, but
  also longer distance
• Trend 2 low cost, low energy, high capacity,
  auto-configured, environmentally hardened access
  packet switches
• Trend 3 ultra high capacity (perabit) backbone
  switches

27
Implications for Network Architecture

• Deep fiber access switch ? large scale packet
  access networks
• In contrast, traditional access network size
  limited by copper transmission distance
• Large scale access network ultra high speed
  backbone switch ? architecture with
• regional node that terminates access network and,
• backbone network with a smaller number of richly
  connected switches

28
An Example Structured 100x100 Network
29
Observations that Lead to Simplifications

• Tremendous focus on protocols/protocol
  architecture, should also focus on
• networks and network architecture
• Interplay between network architecture and
  protocol architecture
• Architecture agnostic to technology trend
• Should leverage technology trends that enable
  simple network architecture
• Box/Switch/Router centric distributed control has
  fundamental limitations, should also explore
• Control architecture that can implement a wide
  range of network-wide policies

30
Robust, Simple Control and Management

• Network is about coordination of switches
• Distributed state management
• Multiple goals
• Reachability
• Policy control
• Security
• Resiliency
• Traffic Engineering, load balancing
• VPN
• Multiple layers (switching optical)
• Diverse switching primitives (OXC, label
  switches, IP switches, vLAN switches)
• Status quo of control and management extreme
  complex, non-linear and fragile

31
Good Abstractions Reduce Complexity
Management Plane
Configs
Decision Plane
Control Plane
FIBs
FIBs
Dissemination
Data Plane
Data Plane

• All decision making logic lifted out of control
  plane
• Eliminates duplicate logic in management plane
• Dissemination plane provides a control channel
  to/from data plane

32
Development of the Elevator Pitch

• Internet technology is far from being adequate to
  support 100x100 vision
• Goal dependable, scaleable, secure, and
  economic sustainable telecommunication
  infrastructure
• It is already extremely complex
• Solution simplify, simply, simply!
• Simplicity should also mean that the design is
  understandable
• Three key ideas that may lead to simplification
• Take advantage of structured network
• Leverage and develop technologies that enable
  structured networks
• Design simple and powerful network-wide control
  abstractions/mechanisms

33
Why 100x100? Why not 1000x100?

• Key is to focus us on an end goal of a network
• Principles derived can be general

34
Why Clean State Design?

• A powerful research methodology that helps to
  crystallize the issues
• Many good examples in systems research that take
  one idea to extreme RISC, SmallTalk, NFS, IP
• A mind set that may result in different research,
  e.g.
• Incremental approach to security
• How to detect and stop Blaster, Code Red?
• Clean state design approach
• What would be the fundamental capability of a
  strategic adversary?
• What are the fundamental limitations/possibilities
  of any network-based or host-based security
  mechanism?
• What should be the minimal necessary set of
  layer 3 security mechanism?
• A concrete and complete different design point
  highlights possibilities
• Understanding the target first helps to plan the
  trajectory of evolution

35
Big Bets Research

• We make big bets to get big breakthroughs
• Visionary ideas carrying intellectual risk
• Cant predict outcomes in advance
• The Christopher Columbus Effect

Randy Bryant Dean of SCS, CMU Strategic Vision
for CS in CMU
36
Can We Make a Difference?

• Monopoly positions in all technology areas
• Microsoft in OS
• Cisco in router
• Intel in processor
• Oracle in database
• People are usually
• too optimistic in prediction of two years out,
  but
• too pessimistic in prediction of five or ten
  years out

37
Can We Change IP and the Internet?

• What is IP anyway?
• Service interface services seen by VoIP, Web
• Data plane (IPv4, IPv6 packet formats)
• Control plane (OSPF, ISIS, BGP, LDP)
• Management plane

38
Learning from Ethernet Evolution Experience
Current Implementations Everything Changed
Except Name and Framing
HUB Switch

• Switched solution
• Little use for collision domains
• 80 of traffic leaves the LAN
• Servers, routers 10 x station speed
• 10/100/1000 Mbps, 10gig coming Copper, Fiber

Router
WAN
Ethernet Conc..
Server
39
How To Evolve?

• One possible path to future
• Fix packet format IP or IPv6 or Ethernet or
  MPLS frame formats
• Evolve switch-switch protocols (NNI)
• control/management plane, where the intelligence
  is

40
(No Transcript)
41
(No Transcript)
42
(No Transcript)
43
(No Transcript)
44
(No Transcript)