Tag Archives: Testing

Information Commissioners View on using Personal data for system testing

Following on from my recent post on “Doing the right thing when testing with production data“, I was discussing my frustation with a colleage at work and they told me to take a look at a copy of the the “Data Protection: Guidelines for the Use of Personal Data in System Testing” document. We had an old copy, and this is a statement from the ICO, in 2003 I believe. There is an updated 2009 version, but I dont have access to this, so I am unable to comment. Either way its a useful snip it to share with everyone.

The Information Commissioner’s view
The ICO advises that the use of personal data for system testing should be avoided. Where there is no practical alternative to using ‘live’ data for this purpose,
systems administrators should develop alternative methods of carrying out system testing. Should the Information Commissioner receive a complaint about
the use of personal data for system testing, his first question to the data controller would be to ask why no alternative to the use of ‘live’ data had been found.
Key risks in system testing There are a number of general risks that exist whenever system testing is undertaken using live data and/or a live environment.

These are as follows:
• unauthorized access to data;
• unauthorized disclosure of data;
• intentional corruption of data;
• unintentional corruption of data;
• compromise of source system data where appropriate;
• loss of data;
• inadequacy of data;
• objections from customers.

There will of course also be sector-specific risks peculiar to each individual business, each type of business and each particular system.
Before commencing any system testing, it is advisable for the data controller to undertake a risk assessment identifying the nature of the risks that apply, their
possible impact and planned handling strategies.

A cautionary tale
The view is sometimes expressed that system testing poses no real data protection problem, as it takes place all the time with little apparent detriment
to individuals. The following case study, which is based on a true complaint received by the Information Commissioner’s Office, shows that the use of ‘live’
data to test systems can indeed cause very real problems for individuals. A pupil was away from home at boarding school. The pupil’s parents received a
letter from the local hospital informing them that their daughter had been involved in a road accident. In fact, there had been no accident, but the hospital
had been using live patient data to test a system for sending out letters to patients.

Do the right thing when testing with production data

I thought I would write a post about organisations not doing the right thing (in my opinion) when they are using production data for carrying out testing. Perhaps I am alone on this one, and I would appreciate any feedback and opinions in the forms of comment.

Now when I talk about production data, I am talking about data that could be considered personal or sensitive data. This could be credit card information, bank details, national insurance number, address, date of birth, medical records, sexual preference, etc. All of this data would be considered highly valuable to a criminal /  fraudster, and as a result should be considered a significant business risk, not to mention the compliance implications regardless of the industry the organisations operate in.

So with this in mind, why is it that so many organisations seem to think that serious consideration doesn’t need to given when it comes to the protection of data when its comes to using production data in a testing environment. Now I fully understand the value of using production data, and the possibly improved quality of testing that can be achieved, however this data shouldn’t be used as is. This data should be removed and sanitised to make it anonymous / de-personalised. This ensures that should the data be compromised, or not cleaned down appropriately or migrated into production this is no real world impact. Also I think we all know that test environments are not often a full representation of a production environment, especially when it comes to security controls.

I have seen this happen in organisations and it can have a real impact on an individual. Put yourself in this situation. Perhaps an organisation who offers health insurance is testing a new premiums engine. They use production data, and kick off testing, changing various parameters, including illness information, and decide to test the impact of having a sexually transmitted disease. Testing then completes, and by some error information is migrated back into production. Next thing the customers knows is they receive a letter saying they can no longer be offered insurance due to being a sufferer of HIV. This information will have been linked with other databases of other organisations, and the domino’s begin to topple.

When we think how something might impact us as an individual we tend to take a little more ownership and care, and I think this is something lacking in alot of organisations. I am not saying creating test data or converting production data for testing purposes is a trivial process, but that doesn’t mean its something that shouldn’t be done. There are various tools and scripts available to do the necessary to production data, and some companies also offer off the shelf test data that may be appropriate.

So next time your involved in testing, make sure you do the right thing. Understand what the goal of testing is, and what the results might look like. Review the data sets that are relevant and the risks and exposures may bring. Then as appropriate do what is needed to mask, scramble, randomise and de-personalise the data. During testing ensure access levels are appropriate, and the necessary logging is in place. Then when all the testing is completed, follow the appropriate steps to clear down the environment ready for next time.

Ideally all this would be clearly defined in security and testing policies and processes, but regardless you will know you are doing the right thing, and this will also help greatly with meeting compliance and regulatory controls. Its probably not considered that this occurs from a consumer level, but doing the right thing could also be considered a marketing benefit.

So rant over, maybe you agree, maybe you don’t, but I would be interested in your comments.