My attempt to get behind the scenes at Epsilon was met with a terse and unhelpful response: "Unfortunately, as we focus on the ongoing investigation, we're unable to comment. Please refer to the statements on our website for the time being."
On its website, Epsilon has the following message for its hapless victims: "Alliance Data Systems Corp. (NYSE: ADS), parent company of Epsilon, today reaffirmed Epsilon's previous statement that the unauthorized entry into an Epsilon email system was limited to email addresses and/or customer names only. No personally identifiable information (PII) was compromised, such as Social Security numbers, credit card numbers or account information."
Much of the discussion around protecting against data breaches has traditionally centered on two important aspects: perimeter security (e.g. firewalls) and data encryption (in situ, and in transit). But there’s another, often overlooked, aspect to protecting your data that’s much less sexy, but no less effective: data architecture.
Data architecture is many things to many people, but typically includes data security (e.g. encryption, addressed above), metadata management, data obfuscation, data modeling, data distribution, and--depending on your perspective--data governance.
How can data architecture help protect your data? Here's a sample series of measures you can take using different components of data architecture.
First, work with your data governance and information security teams to define attribute sensitivity, such as private health information or PII. Update the attributes in your data models to reflect this sensitivity. Then, export this information from your models into your metadata management system, which helps standardize the sensitivity information. Next, propagate it into your other metadata environments, such as your business intelligence tools. Ensure that your analytics and reporting teams are aware of attribute sensitivity when presenting information to users.
Now you'll want to use this information to architect your databases appropriately. Let creative thinking and wisdom guide your data architects and modelers into creating data models that separate sensitive attributes from others. Use query federation techniques in your SQL or application layer to pull this dispersed data together without significant sacrifice in performance. That brings us back to your BI and reporting tools, which is one such place for query federation.
Use data governance policies, driven by common sense, to restrict the proliferation of data across multiple environments. Work with your developer community to define standard operating procedures and techniques, such as data obfuscation that allow for testing application code with "real" data without compromising sensitivity.
Nearly all this falls under the umbrella of "data architecture." And if this sounds like a lot of work in a lot of areas by a lot of people, you're correct. However, you might find solace in the "mathematics of emphasis" philosophy of the late W. Edwards Deming, the guru of quality. It goes as follows: Quality = Results of work efforts/Total costs. So when people and organizations focus primarily on quality, quality tends to increase and costs fall over time. However, when people and organizations focus primarily on costs, costs tend to rise and quality declines over time. Or you could find satisfaction on the immortal words of management and quality consultant, the late Philip Crosby: "Quality is free"--as catchy a phrase as any in the vast world of management theory.
If you haven't given serious consideration to data architecture, now is as good a time as any, because scammers are filling your information aisles with their shopping carts. They'll be paying for your valuable wares with your own credit card. And not just your business, but your job, may well depend on keeping them at bay.
Rajan Chandras has more than 20 years of experience, with a focus on technology strategy, solution architecture and information management. You can reach him at rchandras at gmail dot com.