Defence in Depth. Rings of Security. Multi-layered protection. All well known terms when it comes to protecting information assets. Information can generally be accessed in two ways: via a network or straight from the disk. Organisations pay great attention to policies and controls that help protect information both in transit and in situ. Take a basic network example:
- Company has a firewall configured separating public and private network traffic
- An Intrusion Detection System is also present to detect traffic anomalies
- RADIUS access is managed using two-factor authentication, with tokens
- Within the private network VLAN’s are configured to separate logical business areas
- Physical wired patching is managed and restricted using MAC address tables
- Wireless Access Points have obfuscated SSID’s and complex passwords, with enhanced 256 bit encryption
- Physical machines are managed by group policy with regular patching, local firewalls and anti-virus configured
- Access to local machines is to non-root / Administrator accounts
All pretty standard stuff and most large organisations will probably have securely managed DHCP and DNS too. So that’s the internet to desktop access all sorted, right? Lets look at data in situ. This can be slightly more problematic. Data can be stored in several places: local hard disk, portable hard disk, network storage (server/NAS/SAN) and so on. So more places equals more risk. To overcome some of those obstacles, some basic controls would include:
- Restricted access to physical network storage location (swipe/pin access to server room)
- Locked server cabinets
- RAID storage to reduce SPOF
- Secure back up
- Final resting place disk encryption
- BIOS password to prevent storage changes
- Prevention of USB access to machines holding data
Again all pretty standard and many organisations will utilise some or all of these controls. Individually the controls could be attacked or circumvented, but when combined, the entire chain of security makes an attack less likely to succeed and maybe even less likely to be started due to the formidable steps involved.
The one area that chained security is less effective at preventing is social engineering. The classic fraudster from the American films is not a new concept. Mandrake the master illusionist or the well dressed con-artist who was ‘such a nice man’ all fall in to this category of social engineering.
Social engineering can cover a multitude of areas from tailgating, fake calls from IT support, email phishing right through to elaborate pre-texting and scenario based trapping. All are designed to extract information such as usernames, passwords or personal information that can lead to data leakage. The biggest safe guard for such techniques is often a well developed and disseminated security policy which clearly states, what for example, internal teams would ask for during standard business operations. Many banks now place warnings stating they’d never ‘ask for your PIN during a call’ for example.
The same approach should be standard practice within the corporate environment, with detailed policy explaining that under no scenario would certain information be asked for or released, even if an employee is placed under social pressure. Clearly devised process surrounding information release should be adhered to. If a request needs to be approved by Mr X, allow it to be approved by only Mr X and under no circumstances should that not be the case.
It’s the classic scenario of the corporate directory requiring complex passwords to be used, only for those passwords to be so complex the users cannot remember them so write them under their coffee mat.
Defence in depth is only really as strong as the weakest link and sometimes that is always people.