Big-Data Can Lead to Big Security Risks

Comment | Print | RSS

• Login to Rate

As fields and records continue to grow within databases such as Hadoop (a frequent topic here), things can get overwhelming. At some point, database managers can lose focus and allow human error to kick in. Access to sensitive datasets can easily be granted to the wrong groups of users. That can lead to a downward spiral where confidential data begins leaking out of your massive database without your ever learning about it.

This is especially troubling for companies that are looking to build big-data stacks for highly sensitive subject matter, such as financial information or health records. Smaller databases are easier to manage and have fewer groups of users looking to tap that information. But with big-data, everything is stored under one roof, and there are lots of users who are clamoring to get their hands on all that malleable (and valuable) information.

If they're going to have all this data, it is important for CIOs to concentrate on the basics first, including security. They need to make sure their database administrators are careful to organize the data in a clear and meaningful way. They must always have a sharp eye on the most sensitive pieces of information. They need to know where this data resides at all times and who should have access. Data classification and data flow mappings are crucial, especially from a compliance perspective. It's best that both your database administrators and your security administrators fully understand how these function within the database.

It is also important to monitor and regularly audit data requests from various groups. Ask why a particular user or group needs to be granted access to specific datasets -- or access to the database at all. Databases should be treated with the utmost respect, and that includes strictly limiting access to a narrow group of users. You may end up looking like the bad guy, but at least you'll keep your data safe. And from a career longevity standpoint, it's better to protect your data than it is to make friends with the marketing or sales team.

Before you take your ambitious big-data goals and run with them, take the time to plan out your organization's strategy. Make sure all administrators understand the ground rules and have detailed policies to handle day-to-day operations, as well as security incidents. Believe me -- they will occur. When you're dealing with the amounts of data these databases can collect and manipulate, even the most organized database administrators are likely to make mistakes from time to time.