Paper 5
Layering Public Key Distribution Over Secure DNS using Authenticated Delegation
ACSAC 2005

Why?

To me, key distribution seems to be holding up the whole world of secure networking. I was curious as to how their approach differed from others and whether it seemed to work.

Quick Assessment

Their work looks really nice. I think it is a nice approach to distributing keys.
The premise of the paper is set up in the summary:
Powerful cryptographic tools exist to address security and privacy concerns, but have not been widely used since NO CONVENIENT INFRASTRUCTURES IS AVAILABLE FOR AUTHENTICATED KEY DISTRIBUTION. IKS (Internet Key Service) is a simple, scalable public key distribution service...

Introduction

focus on a capability...simple scalable, authenticated public key distribution...layered on top of Secure DNS (DNSSEC).

Background

public keys must be authenticated to prevent impersonation and man-in-the-middle attacks. Two approaches

IKS uses Certifying Authority.

DNS names are assigned from a hierarchical namespace, and organizations are granted control over a sub-tree rooted at the domain they have registered...

DNSSEC is a collection of proposals for securing the data stored in DNS.

DNSSEC signs the resource records comprising the zone with a public/private key pair bound to that zone, and delivers those signatures to querying clients.

To facilitate distribution of zone keys, DNSSEC defines a DNSKEY resource record. By recursively requesting keys and moving up the DNS hierarchy, the client will ether find a trusted key, or exhaust the name space.

The paper indicates that DNSSEC is now implementable.

BUT

DNSSEC cannot support random key distribution:

Related Work

Can be used to provide a key service:

IKS - Internet Key Service

Design Requirements

IKS Architecture

Protocol Overview


Mike Erlinger

Last Modified Thursday, 30-Mar-2006 09:48:13 PST