There’s data. And then there’s BIG DATA. Many of us have been bombarded with the term in many frameworks. There are some professionals that chalk it up to marketing hype or meaningless buzzword. Personally, I prefer the way Gartner categorizes it. That it is more than size. It is a multi-dimensional model that includes complexity, variety, velocity and, yes, volume.
But the pressing issue with this definition of Big Data is how best to secure something so vast and multifaceted. If you recognize the old concept of a network perimeter is antiquated and dangerously narrow, there should be some concern as to corralling all this data and ensuring its transit and storage is protected. The latter issue speaks directly to compliance needs. Banks and other financial institutions, medical facilities, insurance, retailers and government entities are especially sensitive to the compliance requirements. However, if your business doesn’t fit into these verticals doesn’t mean you can’t directly benefit from cloud based security that creates the necessary context. And though your organization is dealing with an incredible mountain of data, you still must do what you can to ensure not only the proprietary intelligence behind your firewalls, but all the data trafficking in, around and through all various endpoints throughout the enterprise.
But again, size should not be the only consideration regarding Big Data. It is the means by which you analyze and apply various processes that allow you to make the best decisions possible about the ongoing security, accessibility and viability of all those many bits and bytes.
If you are looking at scale the McKinsey Global Institute estimates that “enterprises globally stored more than 7 exabytes of new data on disk drives in 2010, One exabyte of data is the equivalent of more than 4,000 times the information stored in the US Library of Congress. That’s a lot of data.
Storing is one thing, but analyzing and managing all the data into useful strategic and tactical outcomes now depends on the other elements of Big Data (complexity, variety, velocity). To do this successfully you have to have a means to put all of it into context. For instance, let’s say an account is accessed. It has the right user name/password credentialing and seeks to export some personal data or transfer funds, or change sensitive account settings. On its face you should allow this action. They have the right name and authentication. But when this is given greater context, there are dynamics from other silos of information that need to be factored. What is the device profile? URL reputation? Is the IP address consistent? When was last log in attempt? What time did this latest transaction occur? So, what seemed to be a reasonable transaction might shows patterns of anomalous behavior.
But here’s the larger issue—all these factors that play into determining true context (which I call situational awareness) may come from different sources and require a bit of juggling and cross-correlating. You have SIEM, Access Management, Log Management, and Identity Management. And they may all live on various servers in various places within the enterprise. So ensuring this process association is doable, but with so many layers and stacks, the results may take too long to take preventative measures. You know what they say about the horse having already left the barn.
By migrating security functions to the cloud (security-as-a-service) you still may run into these same issues unless you find a provider who can combine all the functionality and create the rules for cross-correlation that can normalize and sort through gargantuan amounts of data. A SIEM solution in the cloud is able to take raw data from a variety of sources, normalize it and create and manage the alerts, escalations and prevention protocols. Such a configuration takes the activity from Identity and access management silos, combines them with the silos of general traffic of web traffic, internal access, SaaS solutions and other business/consumer facing applications and generates a flexible and scalable intrusion detection matrix.
A fully-realized cloud-based SIEM deployment (which is much less expensive in the cloud, yet just as powerful as any on premise solution) can prevent an IP address in China from spoofing your customers account and create intelligence that deflects and notes if a Flame virus is being lobbed at your network. But a true cloud-based security partner worth their salt will also provide the raw data for post-capture analysis. This way you can analyze new traffic patterns, but more important create the baseline to make intelligent decisions for the long term security of your network or immediate recognitions of anomalous behavior. But all that raw data…that’s where the cloud gets you, right? You get penalized for having bigger and bigger data sets. Not if you have the right vendor. I personally know where you can get storage space for as little as $1 per gB per month. You can scale the amount and the type of data you wish to keep in the cloud. You control when it gets destroyed according to various compliance requirements. I also have some thoughts about vendors who provide the services, but require you to buy some appliance that you install and maintain on your network…but that’s a whole other blog.
The bottom line is Big Data can be managed given the right tools. And those tools do exist in the cloud and can be managed through the same. And when you have the right rules, passing though an integrated suite of security solutions you’ll begin to see that size doesn’t matter. What matters is creating a situational awareness that provides you a platform to make better decisions. And if that place is in the cloud…all the better.
Tags: Access Management, Big Data, Cloud, cloud computing, cloud security, CloudAccess, compliance, enterprise-it, Gartner, Identity Management, IT security, Log Management, network perimeter, Security, SIEM