-Of trust: CAs of PKIEP must enjoy of notoriety near
the end-users. Indeed, the governments implication is
more than desirable. Also, it is necessary that the
certificate policies of CAs be available to the certificates
verifiers so that they can judge of the confidence degree
which they will assign to the certificates issued by these
CAs. On this subject, it is very useful that these
policies are written in a formal way in order to allow
an easy and non-ambiguous reading.
- Feasible & convivial: It should not be so much
‘perfect’ at the point to become impracticable. It
should not be forgotten that e-commerce is not
supposed being surer than traditional commerce.
Moreover, if the e-payment applications must become
less convivial to be able to use the PKIEP services,
the users naturally will turn aside from these applications.
These qualities are obviously both general and
informal; they show, however, the great difficulty of
the undertaking task to create this PKI and even the
practical impossibility of this task. I believe, in spite
of that, that while proceeding in a progressive way
this PKIEP-dream can become a reality, especially if
there is behind a real determination and if the
tendency towards the commerce globalization and the
fury towards the e-commerce applications continue.
4.2 Recommendations concerning the
PKIEP characteristics
In this section, it is a question of presenting my
recommendations concerning the PKIEP characteristics
for which I discussed above the general ‘qualities’. The
PKIEP characteristics, that I consider here, relate to its
certificates format, its trust model and its entities.
4.2.1 Certificates format
To remain in conformity with qualities of
universality and globality of the PKIEP, there should
preferably have one format for the certificates.
However, this format must be flexible enough to
contain the various types of certificates, that is to
say, identity and authorization certificates.
X.509 V3 format (Housley, 1999) could be the format
used in PKIEP provided that it undergoes certain
improvements. Among those, I suggest that the
‘name’ field becomes more general in the way that it
will contain information not identifying the certificate
subject, for example, a nickname (
Clarke, 2001).
Moreover, one extension -to be standardized- should
be reserved for the attributes, roles or privileges of
the certificate subject. Indeed, it is sometimes
useless to know the payer identity, but what it is, on
the other hand, necessary, it is to know some ones of
its attributes. I notice besides that the majority of the
individual/payers prefer to keep their anonymity at
least with respect to the paid.
At this level, to minimize the risks of frauds, I propose
that only the CA, which certifies the payer, takes note
of its identity at the time of its first registration. This
CA issues then an attributes certificate to him -after
checking their attributes-which comprises a 'Nickname'
that it will associate to its true identity and this, for
example, in a confidential document that the payer
should sign. Among the attributes which could be
useful to individual/payer, I propose the followings:
age, nationality, profession, police record, existence of
a valid account for e-cheques, e-cash or credit card, a
hash of the account or credit card number, etc.
In addition, I prefer that the paid and the other actors
be identified to avoid many frauds. However, I
suggest also, for the paids/merchants certificates,
that the extension reserved to the attributes contains
information which can be useful for the payers, such
as: Web site address, references, certifications, trading
licence number, jurisdiction, sales turnover, etc.
4.2.2 Trust model and entities of PKIEP
Figure 2 shows the architecture of the trust model
that I suggest for PKIEP. As it is illustrated through
this figure, I suggest the existence of various types of
entities, each one with different functions. Indeed, I
make the distinction, on the one hand, as it is often the
case in a PKI, between two categories of entities: end
entities (EEs) and certification authorities (CAs) and, in
addition, between various types of the same category:
a- EEs of PKIEP:
I suggest making the distinction between two kinds
of EEs: on a side, the web surfers who will play the
role of payers primarily and, more rarely, that of
paids; and on the other side, merchants and
companies present on the Web which will play the
role of paids but also of payers (in the case of B to B).
b- Certification Authorities of PKIEP:
I insist here on the importance owing to the fact that all
the CAs of PKIEP must be trustworthy and especially
‘approved’ by their corresponding governments. Indeed,
it is not necessary that the users of e-payment methods,
all over the world, be constrained to undergo the
monopoly of a private company (like that it seems to be
concretized with Verisign (Verisign, 2004)). In addition,
I suggest that there are various types of CAs in PKIEP:
- PCAs (Policy Authorities Creation): are CAs ables
(and authorizeds) to establish suitable certificates
policies to various contexts of e-payment or e-
commerce. As it is shown on the figure2, I propose that
each country has at least one PCA under the supervision
of the government that can be, in its turn, certified and
E-PAYMENT SECURITY - Recommendations about the use of a PKI for e-payment security
321