Title: A Cryptographic Approach to Safe Inter-domain Traffic Engineering
1A Cryptographic Approach to Safe Inter-domain
Traffic Engineering
- Sridhar Machiraju
- SAHARA Retreat, Summer 2004
2Outline
- Motivation
- Defining the Problem
- Proposed Solution
- Random Noise
- Discussion and Conclusions
3Motivation
- In BGP, Autonomous Systems (ASs) are abstracted
as a node in a graph
4In reality,
Peering links
AS1
Internal links
AS2
AS3
5In BGP,
Peering links
AS1
Internal links
AS2
AS3
6Motivation
- In BGP, Autonomous Systems (ASs) are abstracted
as a node in a graph
- Why?
- Scalability
- Confidentiality of intra-domain information,
e.g., link quality, routing, flow info, policies
etc. - Why is this bad? Traffic engineering by one AS
can send flows over bad paths in neighboring ASs
7Outline
- Motivation
- Defining the Problem
- Proposed Solution
- Random Noise
- Discussion and Conclusions
8High-level Problem Statement
In A, this path has most available bandwidth
A
B
Source of flow F
9High-level Problem Statement
In A, this path has most available bandwidth
path with best end-to-end available bandwidth
B
A
Destination of flow F
Source of flow F
10High-level Problem Statement
In A, this path has most available bandwidth
path with best end-to-end available bandwidth
B
A
Destination of flow F
Source of flow F
- Design a technique so that neighboring domains
conduct traffic engineering cooperatively in a
scalable fashion without having to reveal
confidential intra-domain information?
11Formalizing the Problem
- Consider traffic from A to B that can exit one of
P peering points
- Two kinds of constraints (of A and B)
- Given demand Ti, find amount of traffic, xik of
flow Fi to transit peering point k - For every bottleneck link, , all traffic
traversing it must not exceed avail b/w
Confidential information
12A Linear Programming Problem
Constraints in AS A (private to A)
amount of each flow exchanged at peering points
Constraints in AS B (private to B)
- Objective maximize/minimize CTX
- (minimize) maximum link utilization
- (maximize) total traffic exchanged
- (minimize) average/maximum path inflation
13Outline
- Motivation
- Defining the Problem
- Proposed Solution
- Random Noise
- Discussion and Conclusions
14Overview of Solution
LP1
- Sub-matrices of V,W are private to A, B
- A and B transform the above into
- Solve LP1 and XQX
- V, W, X, X, C, C do not reveal any
information about private information of A and B
to each other (almost)
LP1
15Transforming the LP problem
- A sends encrypted sub-matrix, E(VA) and E(WA) to
B - B chooses random invertible P and Q
- B sends E(V)PE(V)Q and E(W)PE(W)
- requires addition of encrypted values and
multiplication by known scalars (VB, WB) - These can be performed by homomorphic encryption
schemes, e.g., Pailliers - A decrypts E(V) and E(W) to obtain LP1
16The Final Solution
B
E(VA), E(WA)
E(V)PE(V)Q E(W)PE(W)
A
A
Send XQX
Solve VXltW for X
B
E() represents encryption by A
17Outline
- Motivation
- Defining the Problem
- Proposed Solution
- Random Noise
- Discussion and Conclusions
18Small random noise is OK
- LP1 does not leak any information about VB, WB
only if V has full rank - So, add small random noise to matrix entries
- this can be done by homomorphic encryptions
- How does this affect the LP problem?
- Constraints may not be violated by small noise
- Objective function may be affected, though
19Effect of random noise(1)
- 10 constraints objective maximize flow
20Effect of random noise(2)
- Objective maximize (1path inflation)
- About 2-3 unsolvable problems too!
21Outline
- Motivation
- Defining the Problem
- Proposed Solution
- Discussion and Conclusions
- Random Noise
22Discussion
- Scalability
- LP problem transformation is quadratic in terms
of number of cryptographic operations - But, traffic engineering not frequent (hourly)
- Threat model
- ASs are assumed to be rational, i.e., do not
inject wrong inputs - Future work Experiment with real topologies and
quantify time complexity
23Conclusions
- Inter-domain routing could benefit a lot from
cooperation which is hindered by confidentiality
requirements - We demonstrate this for the case of safe traffic
engineering - Other cases of inter-domain cooperation policy
safety, resource allocation and intrusion
detection - checking global invariants
- computing global functions