Prophylactic Data Security?
by TJ Dale - Managing Director of ArmourSoft - Tuesday, 01 July 2003.
The industry is being pulled in two directions - the need to proliferate access to corporate data - everything from global trials to Internet therapy portals - and the need to protect intellectual property rights. The FDA has added its two penn'orth in the form of Part 11, but has chosen (quite rightly, I believe) not to be proscriptive, by simply issuing guideline objectives. So, where does that leave managers tasked with securing corporate and operational data?

It should be simple. First task is to restrict access (using anything from passwords to fingerprint recognition to smart cards); secondly, to frustrate anyone who manages to breach the first hurdle (typically, with encryption techniques). The reality is far from simple. Let us begin by examining some of the exposures.

Biometric watchdogs - without a full set of teeth

Biometric techniques, such as fingerprint recognition or retinal scans are very much flavour of the month and they certainly do have some compelling strengths. They are highly specific and incredibly easy to use - there is nothing to remember. They are also high tech, so they must be good.

In reality, they are nothing more than a complex password. Putting aside the problems and cost of implementing the necessary hardware across the extended enterprise, including field workers such as trials monitors, not to mention CROs, there can still be an overwhelming weak link. If, in order to validate the fingerprint or scan, a record of the original is kept somewhere on the network, then that record is just as vulnerable as any other record on the system. Not only is it vulnerable to external hackers, it is also sitting on a network that is managed and probably routinely backed up by a junior in the IT department - available to be copied electronically. In fact, biometric checks are nothing more than cleverly packaged static passwords and, apart from ease of use, can be almost as vulnerable.

Encryption - a shaky Tower of Babel

The concept of encryption is not new - the Third Reich employed an (almost) perfect technique of converting information into gibberish that could be deciphered only with the appropriate key. Modern computer encryption techniques are remarkably secure, but they can suffer operational problems.

Until recently, the problem with encryption has been the need to compromise - a choice between one of two strategies. Either adding more bolts to the door or accepting enormous overheads in terms of developing, rolling out and maintaining the encryption environment. Adding more bolts (too much authentification) makes for a user un-friendly operating environment, with complex logging on and spending what can seem like an eternity waiting for the on-screen 'egg timer', encouraging users to disable encryption. On the other hand, making encryption transparent to the user (normally through something known as partition encryption) involved high costs, and a protracted period of implementation with, for example 1 gigabyte of disk space taking typically 40 minutes to encrypt, even when empty!

The proliferation of laptop computers also compounds the potential for security breaches. Typically, a user's encryption key has been stored on the laptop (it would be pointless sending it up and down the network). Laptops are frequently stolen from cars - complete with highly sensitive data complete with the encryption key!

Outsourcing the non-core - outsourcing the secrets


The big picture of protecting and securing Big Data

Today almost every company is dealing with big data in one way or another – including customer data, tracking data, and behavioral marketing information – connecting every aspect of our lives. While this is a cutting edge use of technology, data monitoring can become dangerous when placed in the wrong hands.

Weekly newsletter

Reading our newsletter every Monday will keep you up-to-date with security news.

Daily digest

Receive a daily digest of the latest security news.

Fri, Aug 28th