Identity theft is the top security issue facing banks today.
While efforts to tackle fraud have begun to show positive results, for example,
the introduction of EMV saw a reduction of 13% in the first half of 2005 in
the UK according to APACS, identity theft levels are growing. ID theft has
increased by 500% since 1999 and now costs the UK economy £1.3bn a year,
forcing defences against this crime to evolve rapidly. In the past few years
this has led to the emergence of smart card-based integrated authentication
networks.
It is worth pointing out that one of the main reasons that techniques to
defend against identity fraud have not already been widely deployed is the
cost involved. Organisations, be they banks, other enterprises or the public
sector, have struggled to justify the extensive costs that are associated
with mass-deployed and flexible identity systems that are required to meet
the identity fraud challenge. Yet there have been high profile examples
of extremely successful deployments such as the award-winning BACSTEL-IP
system.
Identity security has developed beyond the simplest form of authentication
where one party issues and verifies identities within a closed group of
users. While easy to do, this approach is extremely hard and costly to scale
upwards and offers no interoperability with other authentication networks.
As such, a more common technique being used widely today is the use of
a federated identity network. This allows individuals to use one form of
identity to authenticate themselves to a range of different organisations.
As such, an individual could use one username/password, token/PIN, digital
certificate/passcode issued by one organisation for authenticating themselves
to a completely different organisation.
This approach solves many of the problems associated with the closed group
identity security approach. For a start, the enormous investment involved
in issuing digital certificates on smart cards, for example, can be recouped
to some extent, by deriving revenue from allowing other organisations to
authenticate their users with the same identity. There are also benefits
for the users themselves as they do not need to carry or memorise a wide
range of different passwords, PINs or tokens such as smart cards. In addition,
co-branding and other marketing opportunities can be used to generate extra
revenue within the community of organisations that are using the same federated
identity group.
There are, however, significant challenges with this approach. Central
to this is the level of trust that must be given by an organisation using
another organisation’s authentication network. Essentially, an organisation
that is joining another’s authentication network must have confidence
in the checks that have been carried out to guarantee the identity of the
user. Privacy laws have further compounded this as one organisation is unlikely
to be able to share any meaningful information with another organisation
to prove that these checks are robust.
The result is that such schemes tend to rely upon the lowest common denominator
authentication (typically username/password) and are used within small communities
of interest. The ability to use a federated identity approach in a highly
scalable environment is limited because of the levels of risk involved in
relying upon the work of the issuing organisation. However there have been
notable implementations such as the USA CAC programme where the departments
are required to use a universal registration process.
It is for this reason that a new integrated identity approach has emerged
in the last two to three years. In this approach the focus is on a single
application being used at the hub of the authentication network which allows
all participating organisations to be issuers.
The key lies in the use of an authentication platform that is flexible
enough to accept the digital credentials of any participating organisation.
A good example of this is the new BACSTEL-IP system. When BACS (now VOCA)
wished to deploy a smart card based digital identity solution, it needed
to be able to authenticate a variety of different trust schemes, including
Identrus and various member banks’ own PKI certificates. It therefore
faced head on the need to have one platform that could authenticate any
trust scheme. The success of the system that was eventually deployed has
since received critical acclaim.
An additional advantage of the integrated approach is that it need not
err towards the lowest common denominator digital identity solution – i.e.
username/password. Therefore, should an organisation within the integrated
identity group want to be able to use stronger identity for some, if not
all, of its transactions then this is possible without interfering with
the requirements of other participants. As such, one organisation may consistently
have high transaction values that would justify and require a more robust
authentication solution than lower value transactions would. This is based
upon a financial risk versus cost of solution basis but does allow for the
widespread use of a single smart card-based solution.
Integrated identity also circumvents the data privacy issue as each organisation
need only access the particular information required during the authentication
process for that transaction to take place. If the user wished for personal
information to be held for use by multiple parties then this can still be
done. However, it is much more likely that each application will hold the
minimum amount of information required. Other, more sensitive details such
as trading limits can then be held completely independently of any other
party within the community.
It is envisioned that both federated and integrated identity networks will
be increasingly deployed. The federated approach is thoroughly suitable
for small communities that have rich, intensive interactions with simple
risk management requirements. For large application owners who want a simple
means of deploying and managing applications across a range of communities,
the integrated approach will be preferable. This is especially because it
also allows for a tiered authentication and risk model, and addresses privacy
issues. Furthermore, it has the advantage of allowing any digital identity
credentials that are already deployed to be used.