May 30, 2013

Big Data, IoT, API...Newer technologies protected by older security

 

This guest post, written by Andy Thurai, originally appeared on ProgrammableWeb. Andy is the Chief Architect & Group CTO for the Intel unit responsible for Cloud/ Application security, API, Big Data, SOA and Mobile middleware solutions. You can follow him on Twitter at @AndyThurai.

Nowadays every single CIO, CTO, or business executive that I speak to is captivated by these three new technologies: Big Data, API management and IoTs (Internet of Things). Every single organizational executive that I speak with confirms that they either have current projects that are actively using these technologies, or they are in the planning stages and are about to embark on the mission soon.

Though the underlying need and purpose served are unique to each of these technologies, they all have one thing common. They all necessitate newer security models and security tools to serve any organization well. I will explain that in a bit, but let us see what is the value added by these technologies to any organization:

IoT – is specific data collection points that employ sensors placed anywhere and everywhere. Most often times the information collected by these devices are sensitive data and contain specific identifiable targeted data. IoT allows organizations to analyze behaviors and patterns as needed but also poses an interesting problem. Gone is TB (Terabytes) of data; now we are talking about PB (petabytes) of data which continue to grow exponentially. IoTs use M2M communication, which are a newer channel and create a newer set of threat vectors.

Big Data - - store massive amounts of data (some of these data are from the aforementioned IoTs) and having the necessary software and infrastructure that allow you to access them faster which promises to cost you a fraction of what it is costs today, further enabling you to capture as many data points as possible.

API – interface, enabler and interconnector between systems by providing a uniform and portable interface (whether it is to the big data or the platform that enables big data).

While each of technologies at first glance appears to be serving different constituencies within an Enterprise, there is an undeniable interconnectedness that exists. The IoT collects data from everywhere. Hence, it is pouring tons of data that need to be not only stored somewhere, but also analyzed properly so that the dots can be connected, to ultimately form meaningful patterns that people can make use of.

 With the evolution of these technologies, there is a very raw, basic, and yet incontrovertible need being expressed. Every business yearns to be better than its competitors in catering to the needs of its consumers. I mean the “consumer” in a loose sense here – be that an individual or for that matter, an organization that is consuming your offerings. Ipso facto, this means you need to capture as much information as you possibly can about the target consumer behavior, so that it can be analyzed, protected, stored, shared selectively, and most importantly, so that it can serve your consumer better (or perhaps  to be used when strategically monetizing an area of your business).

None of these technologies is in a trial phase any more. If anything, the social media explosion provided ample evidence that these technologies are being used quite effectively already (real life POCs). Of late, all of these technologies have been gaining adoption in the sacred technology worlds, such as the healthcare and financial sectors.  However, when you employ these technologies with your production applications, you need an enterprise grade security that is built from the ground up to provide a necessary level of protection.

In the social world, the model had always been, “build [it] first and secure later based on the need” (or never in some cases). With healthcare, federal and financial sectors, that model is no longer tenable. You need to secure data at any cost, question anybody who wants access, and be hyper-vigilant without compromise.

What is particularly troublesome is that these organizations seem to be of the thought that they can extend existing security measures to protect all of these newer technologies. While your SSL, Identity systems and other existing controls can serve as the baseline for these technologies, you need a newer set of security controls and tools in place. Your security model needs to make the necessary accommodations, instead of trying to force fit everything to make the older set of tools to fit. That would be like trying to fit a square peg in a round hole. I have seen customers trying to bend RACF to fit the newer SOA, API, Big data paradigm. While it can be done, it would end up costing you more, will be very inflexible, and defeats the fundamental purpose of security. Don’t get me wrong -- everything has a place in this universe.

I recently wrote about the disappearing perimeter defenses and moving lines of thin defense. This is due to shared data centers, cloud adoption, multiple shared tenants, deeper integration and wider exposure to multiple partners, etc. Regardless of the scenario, you need to protect your own data and be accountable for it. Cyber attackers are very sophisticated and are funded by organizations (or even countries), which means they need to get to the proverbial data goldmine.  Without adequate protection, this can prove to be that goldmine. The thing that scares me the most is the underlying threat to all of the above technologies when you try to fit them into the older security model. Most of the above technologies, from what I have observed, are either under protected or unprotected. While it is great for organizations to maximize monetization and satisfaction of a consumer and have a competitive edge over others, that shouldn’t come at the cost of security or by increasing their risk. Especially when it comes to security, Murphy’s Law is always right; it is not a question of if a security loophole will be exploited; it is a question of when.

You not only need to identify the users, authenticate them, and authorize them but also make sure they are allowed access during that time window that they are requesting the info (throw in a location based and device based identification on top).

In addition, you also need to worry about protecting the big data store itself, including strong encryption of storage, transmission, and in process data.

But then, most important of all, you need to mitigate the threat vectors that are created by these new technologies. I will write in the next few articles about how you can protect all of these areas with minimal effort while keeping your TCO very low. I will also talk about specific usecases and usage models that will make sense.