Home Resources Synaptic website articles Security: Information assurance article: Addressing single points of failure in security systems
article: Addressing single points of failure in security systems
Synaptic website articles - Security: Information assurance

Security has often been described as a chain of interconnected links working together.  The well-known industry saying goes that the security of the system is only as strong as the weakest link. These weakest links are single points of potential failure that can result in the security of the entire system failing.

Let us briefly look at one such example. The Kerberos authentication system is designed to allow a trusted third party to securely introduce a user to another user or service. The problem with this system is that the trusted third party is potentially aware of all secret material exchanged between the users. The Kerberos system works well in environments where the trusted third party is also the party providing services to users. But it is not particularly suited to use in other applications. It was exactly this type of problem that led Whitfield Diffe to co-invent public key cryptography:

The fact is that I was counter-culture, so anti-establishment, I could not understand the cryptography in which more than two people knew the key [ed: key translation centres, key distribution centres, Kerberos].

I never understood classical key distribution till much later.

Whitfield Diffie, co-inventor of public key cryptography, speaking at the 30th Anniversary of PKC, Oct 2006

Design of a system with single point of trust can also cause problems when the system is scaled to support a greater number of users and servers.  Read the attack to learn more about the security failures that occur in the Kerberos system when it is extend to rely on multiple single points of trust within a federated system.

While being aware that the system is as strong as the weakest link it has often been used to justify the use of less-than-ideal components and weak constructions on the pessimistic basis that the wider system in which it is implemented is inherently insecure.  The limitation in this line of reasoning is that there is no longer any strong point on which we can manage the points that are at risk.  If our goal is to strive for risk management and the design of high assurance systems than we must design and implement each part of the system in a robust manner. Having comprehensively built these strong foundations we can then appropriately shift our focus to addressing the complex dynamics of large electronic systems and human interaction.

Synaptic is designing a unified ecosystem which addresses the long-term security risks and the point-based problem solving to create a strong foundation on which the perpetual wheels of the security process can stand on.  Synaptic Labs' Group, Enterprise, and Universal key exchanges are exemplary protocols that manage the complex human trust relationships to mitigate against the risk of single point of trust failure. The Synaptic key exchange protocols also demonstrate the ability to layer two fundamentally different key exchange techniques in a manner that provides practical increase in security under some realistic attack scenarios.

Last Updated on Monday, 05 January 2009 15:22