Understanding
WAP Security
On the face of it, the
much hyped WAP appears
W
AP (Wireless Application Protocol) is being much hyped as part of the
mobile computing revolution e business and e commerce computing
over mobile phones. There is certainly going to be a lot of WAP about (IDC
has forecast that all personal communications service phones will be Internet en
quite secure, but there are
abled using WAP by mid 2001), and as WAP is intended to cope with commercial
severa; issues you should
transactions, security will be a requirement.
be aware of.
On the face of it, WAP security appears to be pretty good. Digital phones are trusted
and the WAP model includes security in the form of WTLS (Wireless Transaction
Level Security). WTLS is a form of SSL (Secure Sockets Layer) optimised for mobile
By David Norfolk
phones. Nevertheless, things aren't as simple as this. For a start, some of the threats
IT Journalist
associated with WAP are different to those associated with ordinary Internet
computing (and some are the same), and this must be taken into account when
designing WAP applications. However, security firms such as ISS think that there
are serious and fundamental problems with the WAP security model and, even if
these don't affect you, there are certainly implementation issues with WAP (which
isn't a mature technology) that may compromise security.
The fundamental problem with WAP security is that it is optional. This leads to the
risk that someone can be persuaded to think that a transmission is safe when it isn't,
and this may lead to a further compromise of security. Some of the people involved
in WAP want to change this, but it seems unlikely that this will be addressed even
in the upcoming next release. Perhaps security is seen as a usability barrier to
adoption of a new technology, and many first time apps are simple and don't need
security, but even a perception of a lack of security can kill confidence in a new
technology.
Cryptography
However, the cryptography used for transmission across the Web probably isn't a
problem. No doubt it can be broken, but not easily enough to be generally useful
(that is, not in anything like real time). Where a serious problem does arise is with
the WAP Gateway, which converts from the WAP security protocol WTLS, used
between phone and Gateway, and SSL, used over the Internet. This implies that the
message is available for a short time unencrypted on the Gateway and, if the
Gateway is compromised, so are any WAP communications. Work is under way to
address this (it seems to have no real advantages for anyone), but it isn't known
when, or even if, the WAP specs will change.
A second fundamental problem at the moment, but one that will presumably go
away in time, is the essential immaturity of the technology. Modern mobile phones
are pushing the boundaries of what you can squeeze into a phone (both physically,
and in terms of processor and memory), and this is a classic recipe for the production
of hardware specific solutions, code that is hard to understand, and obscure bugs
introduced in maintenance releases. Tom DeMarco pointed out years ago (in
Controlling Software Projects, Yourdon Press, 1982, ISBN 0 13 171711 1, reporting
work by Weinberg and Schulman in 1974) that giving a programming team a goal
of optimising one project metric (completion time or program size, for instance)
severely impacted performance on other project metrics (program clarity or user
friendly output, perhaps). It is unlikely that the situation will be much different for
the code in a WAP phone (although at least it isn't running Windows applications).
The net result of this technology immaturity is that phones tend to be subtly different
Issue 131:June 2001
PC Network Advisor
File: B1423.1
page 9
Buying and Evaluating:Hardware
www.pcnetworkadvisor.com
Next page >
New! The best sites for quality inkjet printer cartridges and the best sites for cheap inkjet cartridges