The Register has an article that claims that 29% of directors say they steal corporate data when they leave a company. 24% of thefts were done using USB devices (sticks, MP3 players) and 18% used email. There are no excuses for this … this is just plain theft of company data to bring to a competitor so that they have an unfair and probably illegal advantage.
So we’ve identified that USB and Email make up 42% of data theft mechanisms. What do we do? The first thing to do is lock down access to resources. This goes from the basics of controlling data access to controlling device access.
Data access is one of the simplest things to do but is rarely done right. First of, use Active Directory groups to grant access. I can’t think of a place I’ve been to where they haven’t granted access to users directly. That’s just plain dumb and impossible to manage. Next, define owners of the data. This should be a number of people who are in a position to grant and revoke access to data. Only they should give permission to IT to grant access to a user. People automatically assume that IT know who should have access … how can we? Data access is a business issue, not an IT issue. We control the mechanism but not who needs access to the data. Using a strictly enforce and audtiable procedure will control access and give auditors something to track. You can do this with paper but I’d look at a Sharepoint Services site and Infopath (from the Office Professional suite), maybe with a SQL back-end. By tying this with PKI/certificates you can implement a rapid, paperless system with trustworthy signing.
Then there is device access. How many users really need access to a DVD/CD writer, USB sticks, etc, to carry out business? It will be less that 10% in a typical mid sized organisation or larger. For now, the best solution I’ve found is DeviceLock. This service can be installed on all desktops to put permissions on all interface types, e.g. read only CD/DVD, no access to USB, access to USB printers, no access to FireWire. Permissioning is done on a group basis so you can allow local administrators full access, restrict access to all normal users and grant specified access to security groups. For example, I’d have a group called USB-Read and another called USB-Write. The deployment of the agent would configure these groups with the appropriate permissions on every machine on the network (this can be done during install, from a central console or via group policy). Then when a user has a manager state they have a business need for a device, e.g. to write to a memory stick, I’d drop them into the USB write group. Microsoft is promising similar functionality in Windows Vista, managed by Group Policy.
A few years ago I was working in a leading pharmaceuticals site as a consultant. A manager came up to me and asked quietly to investigate something. A sales person with access to sensitive data had left the company to go to a competitor and they suspected that this person had forwarded large amounts of data via email. They asked me what could be done to find out what had happened. I asked them "what auditing have you" and they responded "none". They were $£^& out of luck.
Even with restricted access, it’s possible for someone to steal data. A person with access to company secrets could gain authorised access to a memory stick and everyone has access to email anyway. So auditing is necessary.
Firstly, enable auditing on sensitive resources such as file shares. Make sure you audit successful and failed access. You need to monitor failed attempts but the purpose of this exercise is to monitor theft of data that someone had legitimate access to.
Anyone who has looked at the security event log in Windows knows that you might as well read the Egyptian Book of the Dead … it makes more sense. And what do you do if you have many servers? Are you going to look at the log of every server and trawl through the endless events that pop up for each file access or folder opened? At the moment, you can use a crude tool called EventCombMT. It is pretty crude and sucks to use with servers spread across a WAN. Unix and network types are used to Syslog. There are 3rd party implementations for Windows but here’s the catch. It costs more money and in the end, it’s just copying the noise that is the security log from every server to one point to create an even bigger amount of noise. Microsoft have been working on a solution for years called Audit Collection Services. It’s finally on the way as a part of System Center Operations Manager 2007 (MOM 2007). It will gather key events, soon after they happened, and store them in a central dedicated SQL 2005 database. This database can be secured for auditor access only. It also has a view for reporting so that you have a simple view of the data, presenting the information as if you were browsing the Security Log.
That covers file shares. Next we need to look at email. If this is a worry then you need to implement mail auditing. In fact, in certain regions or industries, you are meant to be doing this already. My experience is that certain regulations such as IFSRA or SOX are being deliberately misinterpreted or ignored so that IT costs can be minimised.
Commvault provides a compliance solution called DataArchiver for Microsoft Exchange. This will capture mail traffic and store it in a secure database that only selected people, e.g. auditors, security officers, IT, can access. This gives you an investigative tool you can utilise to track suspect misuse with.
Your email anti-virus might offer some basic functionality you can use if you don’t need or can’t afford full blown archiving. Microsoft Antigen has the ability not only to filter certain file types but you can capture attachments. A past colleague once caught some nefarious activity with email attachments, something that was strictly banned, by using Sybari Antigen (as it was called then).
At this point , we’ve put all the tool in place. What’s left? Nothing surely, because this is an IT problem, right? Nope. Far from it. Like some sensible security consultants tell us, we can put all the mechanisms in the world in place but in the end, the "meat" will be the weakest link. What do I mean? Humans who want to advance their career or appear helpful will do what ever they can, including contravening procedures and rules.
A while back, I did some work at a finance company. A foreign branch manager had been caught on our proxy logs as a heavy and long term browser of unknown (and hence unfiltered) pornographic sites. We reported this to the the necessary internal authorities but nothing was done. Strange, because 2 other people had been quietly let go for the same actions over a 2 or 3 day period. Then late one Friday evening I’m called into an urgent meeting. The security officer and head of auditing revealed to us that this person had quit with no notice. They suspected this person had burned a large amount of data onto CD. But this shouldn’t have happened because the security officer thought he’d changed this persons access rights. What was the problem in this situation? Firstly, the company turned a blind eye to this persons activities because they were seen as a strategic asset in a new market. When this person quit there was a suspicion there would be a problem but IT was not told. The security officer, who was overr
ated, did not understand how Active Directory worked and had failed to make the necessary changes to restrict access to USB, etc. Had we known, this person who was leaving would have lost all access in a matter of seconds. The IT staff in the branch office were completely unaware and actually granted access to the resources for the leaving manager; in fact it was thought that they even helped with burning data onto CD.
One of my biggest gripes in the corporate world is unequal application of company policies. Internal Audit and Security departments spend the majority of their effort watching and analysing people such as IT administrators when they ignore or turn a knowing blind eye to the activities of their directors. Consider the risks, an IT administrator with access to company secrets knows he’s being watched/audited and won’t take a stupid risk. And the chances of an IT administrator even knowing where to start to look for secrets are minimal. On the other hand, a director or senior staff member knows (a) what secrets there are, (b) where they are kept, (c) has access and (d) no one will even blink if a director shows up in audit logs accessing information … assuming there are logs in the first place!
So what needs to be done? Together, union representatives, security, auditing, IT and solicitors must define policies. These policies should dictate how access is granted and revoked. Unathorised use of data or resources must be defined and prohibited. Punishment must be detailed for contravening these policies. The key component is that the directors must publicly back, enforce and comply with these procedures. A rule is worthless if not applied equally. I dare any HR person to sack an employee for doing something that managers get away with even though procedures ban it. They’ll be in an employment tribunal coughing and bleeding up money in a very public and embarrassing manner.
- Control access to data.
- Restrict access to resources, e.g. USB, CDRW, etc.
- Audit and track usage and communication of data.
- Clearly define and communicate policies. Equally and fairly enforce the policies.
This blog post is the property of Aidan Finn (@joe_elway / http://www.aidanfinn.com) and may not be reused in any manner without prior consent of Aidan Finn. You may quote one paragraph from this blog post if you link to the original blog post.