Security can be simple and secure
It
was interesting to read the press release given by Phil Dunkelberger,
CEO of the newly formed PGP Corporation, saying, “You are faced with
the situation where usability is traded off against security - the more
usable something is, the less secure it is." But why must
security be difficult to use in order to be secure? Well the
reality is that it doesn’t! However, if you wanted to believe the press
release then you could be forgiven for mistakenly believing that it
must be so.
Let us use some logic here to analyze the thinking.
If it were true, then anything that is simple to use cannot be
secure. The fact of the matter (as the lawyers would say) is that it is
not essential to make something difficult to use in order to make it
secure. Rather the opposite.
There is no reason at
all why something MUST be difficult to use for it to be safe and
secure. Look no further than your automobile vendor. Do you
suppose that they make cars more difficult to use so that they are more
secure? Get real!
So why have we got this strange idea of` "No pain, no gain?"
The
basic problem seems to start with the IS security community themselves.
In order to achieve recognition they have to carve out their own
niche that demonstrates how important they are (and why you should pay
their costs). To make yourself important in a technical subject
you have to be difficult to understand so that people cannot readily
question what you do or why you do it. To reinforce that notion
you make people learn all about your own ideas (worship at your shrine)
so that they become true believers.
Quite frankly, if you can’t
understand what someone is telling you, the problem is with them, not
with you. That is not to say that the subject of IS security is not
complex. But so is the subject of how car tires are safe, and
few, if any of us know (or really care) how they are made - providing
they perform.
So that’s the acid test. Can the person
using the security system make it work without too much difficulty?
Well, according to the pundits of the security industry,
apparently not. It seems that the average user (or senior
executive) can’t use security products because they are just too
difficult. Well they probably haven’t tried products from
ArticSoft which take the mystery out of security.
Certainly the
‘Public Key Infrastructure’ which was supposed to solve all our
Internet security problems is very complicated. It is
particularly complex if you look at the web sites of VeriSign or RSA.
It would be a lot simpler if they talked in plain English instead
of ‘Pompous Techspeak’. Perhaps that is one of the reasons why it seems
just too difficult to use.
If you want to talk to someone on the
phone you need their phone number. So if you want to speak to
someone with privacy you need their public cryptographic key (phone
number?). That’s all there is to it. If you don’t happen to know
their number or have it in your diary (keystore, wallet, or whatever
it’s called) then you either get them to tell you (send it by e-mail)
what it is, or you look it up in a Directory if there happens to be one
handy.
If you want to send something that’s really ‘private’
you’re hardly going to do that with someone you don’t actually know.
That would be really silly. You only ever send secrets to people you
actually know! And a secret is hardly going to be sent to
‘everybody’ because that’s an oxymoron (a contradiction in terms
according to the Oxford English Dictionary). So there’s little
point having a ‘solution’ to a problem you don’t actually have.
So
where does that leave us? Well, security products only have to be
complex if the designers make them so. You see, that’s the issue.
Whilst there are too many people who think that users should ‘take
responsibility’ for security by making what are, in reality, arbitrary
choices (what algorithm, key length, password length, quality of
mother’s maiden name, smart card and so on) then security will continue
to wander in the wilderness.
It is a simple matter of
making the presentation of security to the user much simpler by hiding
the complexity and using ‘best of breed’ methods so that the user can
make useful business decisions such as who to send secrets to, secure
in the knowledge (as with the car tire) that the technical ‘stuff’ has
been dealt with by people who know what they are doing.
As a
final thought, if you want your computer to be totally secure, then the
easiest approach is probably to bury it underground encased in
concrete. Plenty of pundits will tell you that you should not use
a system that is not totally secure. Follow that kind of advice
and you will be unable to travel, live in a house, eat or maybe even
breathe. We all know that security has to be good enough provided that
it’s the best that’s available without being impossible to use.
|