Migrate: An End-to-end Mobility Architecture for the NGI

Preview:

Citation preview

Migrate: An End-to-end Mobility Architecture for the NGI

Hari BalakrishnanAlex Snoeren Jon Salz Dave Andersen

Frans Kaashoek

MIT Lab for Computer Sciencehttp://nms.lcs.mit.edu/

Networks and Mobile Systems

iNAT Project: Motivation

• Increasing heterogeneity in the Internet Nodes: Mobiles, devices, sensors,... Links: Optical, wireless,... Services & applications: Web, telepresence,

streaming, remote device control

• Need a general solution for applications to deal with mobility and discover resources

• Need a general framework for learning about and adapting to changing network conditions

iNAT Approach

• Intelligent naming and mobility Resource discovery: Intentional Naming System (INS) using

expressive names and self-configuring name resolver overlay network

Migrate: Secure connection migration + dynamic naming (this talk) Demo this evening: “vertical mobility” between 802.11 & Bluetooth

(Snoeren)

• Adaptive transmission End-system congestion management and adaptation framework

for the NGI Congestion Manager software and algorithms Demo this evening: Adaptive MPEG-4 delivery using CM & SR-

RTP protocol (Feamster & Goraczko)

Migrate

• Mobility support for NGI• Goals

Handle moving nodes Handle “vertical mobility” across network technologies Handle moving applications (across hosts) Handle disconnections followed by reconnection (on-going;

Snoeren’s dissertation)

• End-to-end mobility system Architecture leverages dynamic DNS for host location Secure transport-later connection migration enables

separation of TCP connections from IP addresses

Anatomy of a Connection

• Connection defined by IPA:PortAIPB:PortB

• An IP address does not identify a host; it only identifies a network interface

• Is this a good definition of a connection?

Some socket onnms.lcs.mit.edu

www.cnn.com’sHTTP service (port)

18.31.0.82:5678 207.25.71.23:80

Problem: Host Mobility

• Cerf’s comment from DoD Internet paper:“If a host were to move, its network (and host) addresses would change and this would affect the connection identifiers used by the TCP. This is rather like a problem called "dynamic reconnection" which has plagued network designers since the inception of the ARPANET project in 1968.”

• Problem: Changing IP breaks TCP connections• Two options today for connections:

Terminate and retry Somehow preserve IP address and continue

• Another problem: “Horizontal” mobility isn’t quite enough...

Vertical mobility: Seamless inter-network movement

In-building & In-room

Campus-Area Packet Radio

Metro-Area

Regional-AreaExample: cellular telephonyacross wide-area cellular,802.11 and BluetoothMotivations:• Energy• Capacity

Possible solutions

1. Force constant IP address for end-point Mobile IP Layer-N switches with “Virtual IPs”

2. Make names routable All packets identify destination by name, which serves as

routing identifier INS approach (late binding), also in TRIAD

3. In-band migration Don’t confound end-point and routing identifiers! DNS name identifies end-point; IP address is routing ID

Home AgentIntercepts

pkts

Address Constancy: Mobile IP

Foreign Agent (FA)

Mobile Host

Correspondent Hosts

Temporary address dtmp

changes with mobility

D

“Tunnel” to FA

D

dFA

“Detunnel” to Don addr dtmp

D

D

The Case Against Mobile IP

• Requires additional network support and infrastructure (HA, FA, authentication,…)

• Triangle routing even for “local” interactions• Many types of mobile applications

Connections that don’t care for seamlessness Connection initiators Both initiators and responders

• Ingress filtering requires reverse tunneling too• Vertical mobility can’t be properly handled in

application-specific manner• Applications can’t be made aware of mobility

What Should a Connection Be?

• Between communicating applications, not network interfaces

• Should be possible for an application to easily change network interface of connection: While preserving good unicast routes Securely

• Should not require a priori knowledge of valid network interfaces

• Dynamism should not affect semantics or correctness, nor worsen reliability (or performance)

Migrate Overview

Name server

Migrating Hostfoo.bar.edu

Location Query(DNS Lookup)

Connection Initiation

Location Update(Dynamic DNS Update)

Connection MigrationSecured using Diffie-Hellman

exchange (with SYN exch)

xxx.xxx.xxx.xxxyyy.yyy.yyy.yyy

FixedHost

Problems

• Consistency of name mapping

• Correctness Handling packet losses around time of movement What if someone else gets your old address?

• Security Connection hijacking Denial-of-service protection

• Semantics How to maintain semantics of connection sequence

across different nodes?

Dynamic Name Mappings

• Zero TTL on A-records for migratable names• Use Dynamic DNS (RFC 2136, 2137) for Internet

names• Potential problems:

Race condition between name update and movement• Simply retry! This isn’t a new failure mode

What about extra DNS load?• Turns out low-TTL A-record appear not to affect DNS

scalability much• (See Jung, Sit, Balakrishnan, Morris; “DNS Performance

and the Effectiveness of Caching,” ACM SIGCOMM Internet Measurement Workshop, Nov. 2001.)

Migrating a Connection

• Initiate migration from new network address Identify previous connection with token, on SYN Secure token to protect against hijacking Requires some state machine changes to guarantee

correctness

• Preserves service model to application

• Handles “middle boxes” Works with most NATs, PEPs, stateful firewalls

• Requires changes to transport protocol Kernel TCP, SCTP, RTP (linked library)

TCP ConnectionMigration

1. Initial SYN

2. SYN/ACK

3. ACK (with data)

4. Normal data transfer

5. Migrate SYN

6. Migrate SYN/ACK

7. ACK (with data)

TCP ConnectionMigration

1. Initial SYN

2. SYN/ACK

3. ACK (with data)

4. Normal data transfer

5. Migrate SYN

6. Migrate SYN/ACK

7. ACK (with data)

TCP ConnectionMigration

1. Initial SYN

2. SYN/ACK

3. ACK (with data)

4. Normal data transfer

5. Migrate SYN

6. Migrate SYN/ACK

7. ACK (with data)

Modified TCP State Machine

MIGRATE_WAIT2MSL timeout

recv

: SY

N (

mig

rate

T, R

)se

nd: S

YN

, AC

K

recv:

RST

appl:

migrate

send:

SYN (migr

ate T, R

)recv: SYN (migrate T, R)

send: SYN, ACK

• 2 new transitions between existing states

- and -• 1 new state

handles potential race condition due to rapid readdressing

Securing the Migration

• Problem: Increased vulnerability to hijacking Ingress filtering (RFC 2827) doesn’t help Attacker only needs token and sequence space

• Solution: Keep the token secret Negotiate it using Diffie-Hellman exchange (Elliptic-Curve

DH) Use sequence numbers to prevent replay Crypto exchange completed in SYN handshake Implemented in a new TCP option

• Result: Connections are as secure as standard TCP

Migration Trace with SACK

SYN/ACK

Migrate SYN

BufferedPackets

(old address)

ACKw/SACK

Summary

• The Migrate architecture enables end-to-end connections to be separated from, and move between, IP addresses

• Handles host and application mobility in an end-to-end fashion

• Exports API for mobility adaptation• Current work: handling disconnections• iNAT demo this evening (Bluetooth+802.11 end-to-

end migration)• Got code?

Migrate code for Linux 2.2 available from:

http://nms.lcs.mit.edu/software/

Migrate project Web page (has code, papers):

http://nms.lcs.mit.edu/migrate/

Networks and Mobile Systems