Transcript PPT

15-441: Computer Networking
Lecture 26: Networking Future
Overview
• Learning From Failures
• Changes in Various Layers
• New Services
• What Do I Work On?
Lecture 26: 12-06-01
2
Learning From Failures
• Past failures
• Multicast
• QoS
• MobileIP
Lecture 26: 12-06-01
3
Why Did They Fail?
• Scalability problems
• Incremental deployment
• Interfacing with applications/Building useful
services
• Debugging problems
• Conservative network administrators
Lecture 26: 12-06-01
4
What Can We Learn?
• Avoid same pitfalls
• Clever techniques
• Fair queuing, announce/suppress protocols,
tunneling/encapsulation, etc.
Lecture 26: 12-06-01
5
Overview
• Learning From Failures
• Changes in Various Layers
• New Services
• What Do I Work On?
Lecture 26: 12-06-01
6
Link Layer
• Optical links
• Multiple wavelengths on a single fiber (WDM)
• MPLS applied to wavelengths  MPλS
• No longer broadcast
• All optical networks
• No buffering!!  How does this affect other
protocols
• Mobile/wireless links
Lecture 26: 12-06-01
7
Overlay Routing
• Basic idea:
• Treat multiple hops through IP network as one hop in
overlay network
• Run routing protocol on overlay nodes
• Why?
• For performance – can run more clever protocol on
overlay
• For efficiency – can make core routers very simple
• For functionality – can provide new features such as
multicast, active processing, IPv6
Lecture 26: 12-06-01
8
IP Multicast
Gatech
Stanford
CMU
Berkeley
Key Architectural Decision:
Add support for multicast in IP layer
Lecture 26: 12-06-01
9
Overlay Multicast
CMU
Gatech
Stan-LAN
Stanford
Stan-Modem
Berk1
Berkeley
Overlay Tree
Berk2
Stan-LAN
Gatech
Stan-Modem
CMU
Berk1
Lecture 26: 12-06-01
Berk2
10
Overlay Challenges
• “Routers” no longer have complete
knowledge about link they are responsible
for
• How do you build efficient overlay
• Probably don’t want all N2 links – which links to
create?
• Without direct knowledge of underlying
topology how to know what’s nearby and what
is efficient?
Lecture 26: 12-06-01
11
Congestion Control
• Is AIMD the right choice for everyone?
• What are the requirements on choices  TCPfriendliness
• Non-linear controls
• Rate-based controls
• Fixing poor interaction with HTTP
Lecture 26: 12-06-01
12
Denial of Service
• Objective of attack: make a service unusable,
usually by overloading the server or network
• Example: SYN flooding attack
• Send SYN packets with bogus source address
• Server responds with SYNACK keeps state about TCP
half-open connection
• Eventually server memory is exhausted with this state
• Solution: SYN cookies – make the SYNACK contents
purely a function of SYN contents, therefore, it can be
recomputed on reception of next ACK
• More recent attacks have used bandwidth floods
• How do we stop these?
Lecture 26: 12-06-01
13
Bandwidth DoS Attacks
• Possible solutions
• Ingress filtering – examine packets to identify bogus
source addresses
• Link testing – how routers either explicitly identify which
hops are involved in attack or use controlled flooding
and a network map to perturb attack traffic
• Logging – log packets at key routers and post-process
to identify attacker’s path
• ICMP traceback – sample occasional packets and copy
path info into special ICMP messages
• IP traceback
Lecture 26: 12-06-01
14
Overview
• Learning From Failures
• Changes in Various Layers
• New Services
• What Do I Work On?
Lecture 26: 12-06-01
15
Network Location Service
• Desirable to lookup performance between hosts
• Why?
• How to predict?
• Based on historical measurements
• Based on on-demand probing
• What exactly is performance?
• Bandwidth
• Delay
• Application response
Lecture 26: 12-06-01
16
Services For Mobile Users
• Why?
• (Example) Mobile users are more likely to
search for services near them
• Not well suited to administratively organized Internet
systems
• Example
• Build a wide area service discovery that can
support multiple search styles
Lecture 26: 12-06-01
17
Overview
• Learning From Failures
• Changes in Various Layers
• New Services
• What Do I Work On?
Lecture 26: 12-06-01
18
Three Project Areas
• Congestion Control
• Solving interaction between HTTP and TCP
• Using congestion control to implement QoS
• Mobile Networking
• Making protocols adapt to dynamic conditions
• Helping “ubiquitous” networks evolve
• Sensor networks
• Wide-Area Distributed Applications
• Tools to help developers build large distributed
applications
• Overlay multicast
Lecture 26: 12-06-01
19