By Sarah Gordon
E-mail:[email protected]
This paper was first prepared for the 6 th Annual Executive Information Security Risk Management Symposium, November 8, 1995. © 1996 The Alea Group. This document may not be reproduced in whole or in
part, stored on any electronic information system, or otherwise be made
available without prior express written consent of the author
and publishers.
- Introduction
- Using Bad Sources
- Using Things the Wrong Way
- Conclusion
- About the Author
Introduction
"I have seen the enemy and he looks a lot like me!"
The Internet is many things to many people. It is variously a place to do
research, conduct business, advertise products and discuss issues. It has
as many uses as it has users. But for the unwary it also has pitfalls and
hidden traps. Some of the hazards are the fault of faults in software,
and some are due to poor administration. However, some of the most common
security problems are caused by things the we do ourselves. In this paper
we will examine some of those problems and see how many of them can be
avoided by application of a few basic rules.
The self-inflicted security exposures we will be examining fall into two
categories: using the wrong sources and using the right things the wrong way.
We will look at each of these areas and see how users can be led astray, and
we will give some advice on how minimizing exposure by doing things the right
way.
Using the Wrong Sources
The Internet has been described as the world's only functional anarchy.
Where the "big three" standalone networks many users meet during their
introduction to the Internet (America OnLine, CompuServe and Prodigy) work
on a structure that is imposed from the top down [central administration,
uniform policies and official helpers], the Internet's structure comes from
the bottom up. If someone wants to call himself an expert, there's nobody to
stop him. There's no test to take, no authority to certify him.
There are many good, intelligent and informed people who give of their
time to help others. But there are also many who are not so good, not so
intelligent or not so well informed that also do the same. The result is
that there's a lot of bad advice out there. Users following the advice
they get from these sources can find themselves in worse trouble than when
they started.
Inexpert Experts
The first type of bad advice comes from people who think they know what
they're talking about but don't. Whether it's fixing cars, baking cookies
or securing computers, people like to give advice. Some of the people who
give this advice are experienced and knowledgeable, but some are the ones
who should be listening to others instead. Ask how to get rid of a virus
and you'll get five answers, four of them wrong. Witness this recent
recent exchange on the alt.comp.virus newsgroup (names have been changed to avoid embarrassing anybody further):
From: [email protected] (User)
Subject: Virus from CD-ROM
Hello all, I have a question about viruses. Can I get a virus from a
CD-ROM?
From: [email protected] (Bad Advice 1)
Subject: re: Virus from CD-ROM
Hi User. No, you can't get a virus from a CD-ROM. CD-ROMs are
read-only, so a virus can't get on them in the first place.
From: [email protected] (Bad Advice 2)
Subject: re: Virus from CD-ROM
User: It's almost impossible to get a virus from a CD-ROM. No
professional CD-ROM producer would let a CD be released with viruses on
them. They do many checks on them, more than for a floppy. You can
quit worrying.
From: [email protected] (Bad Advice 3)
Subject: re: Virus from CD-ROM
People worry too much about viruses! It's all hype from the AV
community, who're just trying to sell more software. Viruses don't hurt
anybody; even if you do get one, just use a free virus fixer and get rid
of it. Geez.
If User listens to the advice from these three "experts", he could be
setting himself up for a fall, as there have been several cases of CD-ROMs
being released with viruses on them. Viruses can and do cause real
damage, inconvenience and loss to individual users and companies.
The lesson to be learned from this is that not all advice is equally good;
just because someone sounds like he knows what he's saying or believes what he's
saying strongly, doesn't mean that his advice is worth taking.
Rumor Central
The second type of bad advice is the persistent rumor. These are
constantly floating around the Net. There are several rumors that simply
refuse to die, like the story of Craig Shergold, the boy who wants to
break the world's record for the most postcards. A new rumor surfaced
late last year, about something called the "Good Times Virus". Here is
the original message:
---------------------------------------------------------------------------
| Here is some important information. Beware of a file called Goodtimes. |
| |
| Happy Chanukah everyone, and be careful out there. There is a virus on |
| America Online being sent by E-Mail. If you get anything called "Good |
| Times", DON'T read it or download it. It is a virus that will erase your |
| hard drive. Forward this to all your friends. It may help them a lot. |
---------------------------------------------------------------------------
There is no such thing as the Good Times virus as described here, and when
the rumor started there was no virus in the wild which could infect email
files (this was well before the discovery of Word document viruses). Yet
this notice was posted and reposted to every forum imaginable, on the mere
strength of rumor. People never stopped to check out whether the story
was true, they simply passed it on. Please note the lack of
substantiating evidence in the Good Times rumor; no authority is cited, no
expert is quoted. This is the distinguishing characteristic of all such
rumors.
Heeding unsubstantiated rumors like this one will either end up with the
user feeling foolish at best, or doing something dangerous at worst.
Malicious Helpers
Sometimes the person offering to help you isn't merely clueless or
misinformed, but is actively malicious. Yes, it's shocking but true.
There are some people out there who don't have your best interests at
heart. These are not likely to be found posting in open forums like
mailing lists and newsgroups, but are more often lurking on chat systems
like IRC or contacting users directly.
Some malicious "helpers" only do their damage when they see a likely
target, seizing the opportunity as it comes. They will wait for a user to
ask for help, then give advice meant to harm them, either by destroying
their files or giving themselves access to the user's account. Typically
they will ask the user to execute some arcane command, such as
'/exec echo
"+ +" >.rhosts'
on IRC. This would open the user's account up to anybody
who wanted to rlogin or rsh in from anywhere on the Net.
Some, however, are not content to wait for their prey to come to them.
They seek victims out, using whatever method they think they can get away
with. Perhaps they will send forged mail from the administrator, or send
a spoofed wall message to all users on the host, asking them to change
their passwords to a certain string. Here is an example of such a
message, using a hacker tool that spoofs wall messages:
Broadcast message from [email protected]...
Hello everybody. We're experiencing some problems with the password
file (it got corrupted), and need you to help us out by changing your
password to "iamlame". Wait 10 minutes and then change it back to what
it was, and everything should be fixed. Thanks!
Admin
In all these cases, the predatory hacker is taking advantage of the
willingness of the user to go along with what he's told, and not question
whether what he's doing is wise or reasonable.
Knowing Who to Trust
With all this bad information going around, how can a user know who to
trust? Well, the Internet may be somewhat of an anarchy, but it's not
wholely without order. The first place a user should go is his local
administrator. He knows the system, knows its ins and outs. If he has
problems he knows who he can turn to for answers.
For matters less localized, there's CERT. They have FAQs, advisories and
security tools for every occasion. Just point your web browser at
www.cert.org.
Using Things the Wrong Way
The second major class of mistakes users make is using things the wrong
way. This covers a lot of territory; administrators and system experts
abound with "stupid user" stories. In this paper we'll focus in on one
aspect of this problem, ways users misuse their system that make them more
vulnerable to attack by hackers.
Users open themselves up to hackers in three basic areas. They use
insecure channels for private acts, they extend trust too broadly, and
they announce their vulnerability to the world.
Private Acts on Open Channels
The past few years have seen an explosion in the ways hackers spy on
users. They "sniff" passwords of any user going through a system. They
monitor the sessions of users, seeing everything the user does. They
replace programs that are intended to increase security with versions
that render them useless. Users need to be more aware than ever before
of just how open the Internet really is.
Sniffers
One of the biggest innovations in hacker technology is the password
sniffer. The idea is quite simple. Data on the Internet is broken down
into packets, which are sent from one system to the next until they reach
their destination. Anyone who can sit on a system anywhere along the
route the packets take can sift through them looking for ones that
interest him. When a user connects to a system, the first thing he must
do is log in. The name of the host he's connecting to, his account name
and his password are all sent in the first few packets. All the hacker
needs to do is grab those first few packets and he has everything he needs
to access the account.
If the users are only using their local accounts this is no problem;
since the hacker needs root access to do his sniffing, he doesn't really
need the passwords of local users. But many users have multiple accounts,
and frequently go from one to the other as needed. This means that if the
hacker and user are on host A, and the user telnets to host B, the hacker
(using his sniffer) now has the user's password to host B. Now that he's
on host B he can try to get root and set up a sniffer there too.
The hyper-connected nature of the Internet makes it very convenient for
users to hop from system to system. But many of them never realize
they're bringing along an uninvited guest.
There are ways to deal with the problem of sniffers. First, try to
minimize your exposure by not telnetting from an insecure system. If you
absolutely have to, login as soon as you can from a more secure method
(dialup if possible) and change your password. If security is a major
concern, see if your system has s/key; if not, get your administrator to
install it. S/key is a one-time password system. That means that it
doesn't matter if a hacker gets your password, because it's useless the
second after you've typed it. It's somewhat of a bother to use, but if
you do a lot of connecting from "dirty" sites, look into it. S/key can
be found at ftp://first.org/software/skey.
Snooping Sessions
The next logical step for a hacker, once he can sniff the start of a
session, is to watch the whole thing. This is a fairly recent
development, at least for hackers. Put simply, a hacker with root on the
same system (or even just a system on the same ethernet) as the user can
not only sniff his passwords and read all his files and email, but can
watch his whole session just as if he were sitting next to the user. This
ability is not at all the fault of the user; however, anything
confidential the user types out and the hacker sees, is. This is not to
minimize the moral irresponsibility of the hacker or engage in a game of
"blame the victim", but to alert users to exercize "due caution" online.
If he discusses the unnanounced merger with another firm in a talk
session, he should understand the risk involved.
Thankfully, there is a solution to the problem of session snooping. And
it's not even hard to deal with. The solution is end-to-end encryption.
There are a few software packages that handle this; one of the better ones
is called deslogin. Of course you need to make sure the machines at both
ends of the session are clean, or else the encryption is just for show.
Deslogin can be found at ftp://uu.uu.net/pub/security/des.
Logging Keys
The final area users often fail to take into account when dealing with the
openness of the Net is how they manage encryption. PGP is fast becoming a
standard on the Internet, and more and more credit card commerce is being
done on the Web every day. If a user relies on the security of his
service provider to protect his PGP key and passphrase, he might as well
not bother encrypting it at all. Trojanizing an ISP's copy of PGP to log
passphrases is trivial for any experienced hacker. Doing the same for
credit cards on a web browser or HTTP daemon is no tougher. Encryption
systems like SSL only secure the path, not the endpoints.
If it's worth encrypting, it's worth doing it right. Always use your
home PC for PGP. You're frankly better off doing your credit card
purchases offline for now.
Extending Unwarranted Trust
As the Internet grows more and more connected every day, the line between
the local system and remote ones begins to blur. Much of this is not
under the control of the user, but some of it is. Files that make
accessing remote systems easier abound on the Net. For Unix shell users
there's the .rhosts file. For ftp users there's .netrc. For X-Windows
there's .xhosts. All of these help make the user's environment seamless,
adding considerable convenience and power. But for too many users this
comes at a price, security. A user that puts a "+ +" in his .rhosts or
.xhosts file might as well post his password to alt.2600.
There are other extensions of trust users give unknowingly; ftp'ing the
latest IRC client, not knowing that it has a trojan inside it; changing
permissions on their home directory to allow others to access their home
page or their .plan file, not thinking about all the other files they're
also exposing. Users who assume that security is only a concern for
administrators may soon find out how wrong they are.
There is no simple answer to all the vulnerabilities discussed here.
Users need to carefully examine the decisions they make with their
accounts, and keep their eyes out for anything unusual. Things to watch
include strange files, odd behavior at login or logout, and especially
notice what the system tells you your last login time was.
Broadcasting Victimhood
Sharks can smell blood in the water miles away. Hackers can smell newbies
on the Net half a world away. They give themselves away in so many ways.
They ask questions that reveal far more about themselves than they intend,
things that could be used to compromise their accounts. A question on a
security newsgroup about how to choose a good password tells the hacker
that if he acts quickly, he can probably guess the password in just a few
tries. A request on IRC for the latest toolkit scripts could leave the
user with a script that does more than he asked for.
Users need to use their heads and think before they act. Try to find the
answer to the question on your own, or turn to someone you trust, like
your network or system administrator.
Conclusion
Users can be their own worst enemy. Listening to bad advice, trusting
the wrong people and treating their accounts and programs the wrong way
will surely trip them up in the end. But they can also be their own best
friend, if they use their head, get good advice and use things the way
they were meant to be used.
About the Author