The Importance Of Data Management

The Importance Of Data Management:

Data Is The Key

As many of you know or might not know, I also work during the day time as a content writer. I write on all sorts of stuff, ranging from IT topics the cloud, virtualization, VoIP, security, networking, to even further flung subjects such as hydraulic hoses, belt drive design, to GPS tracking. But in all of this writing, there is one key area which is crucial-the data that drives this stuff, and how the businesses that own this content manage this data. Really, data management is very crucial to any business/organization, no matter how large or how small. After all, you need sales reps to get you customers and revenues, and they need this data so they can find those sales prospects.

Even in the Biometrics Industry, data and data management is also critical. The outsider does not see this, but to us on the inside, it is very important. But in our field, our data is unique-in that it is not just all about numbers, but it is about the biometric templates and the mathematical algorithms which support them.

So, as you can guess today’s topic, here we go: “Neurotechnology, a provider of high-precision biometric identification technologies, today announced the release of MegaMatcher Accelerator 4.0, the latest update of the company's multi-biometric software and hardware solution for high-volume, high-speed fingerprint and iris identification. MegaMatcher Accelerator 4.0 includes a larger database capacity, enhanced fault tolerance and a new matching algorithm that provides even greater accuracy, with extremely fast matching speeds of up to 100 million fingerprints per second – with a tolerance of 360 degrees of fingerprint rotation – and up to 200 million irises per second. Even at these high speeds, MegaMatcher Accelerator performs a full database search for each match (unlike systems that use internal database indexing or pre-classification); this eliminates potential false rejections from incorrect classifications and gives MegaMatcher Accelerator a very high degree of reliability . . . MegaMatcher Accelerator 4.0 provides fast, accurate multi-biometric identification using both fingerprint and iris modes together, or either mode can be used as a primary form of identification. MegaMatcher Accelerator also can check identification results with other biometric data from any Neurotechnology-supported modality, including face, palmprint or voice identification. It can manage a database of up to 40 million fingerprints and 50 million irises with the Extended version and up to 4 million fingerprints and 5 million irises with the Standard version. For even higher volume applications, multiple MegaMatcher Accelerators can be connected in a cluster to manage databases with billions of records.” (SOURCE: http://www.findbiometrics.com/industry-news/i/9346/).

My Take

Now, one thing I want to make clear, I am not trying to make a sales push to any of my readers on this. Rather, these statistics exemplify (as can be seen in the above quote) just how crucial biometric template management really is. Just as much as the hardware is important, so are the databases and the data which resides in them. It all comes down to how quickly one can be verified and/or identified in just a short period of time. It can be likened to waiting in the grocery shopping line. We never want to wait for a long time, and we don’t want to wait for a long time to be verified. Now of course, it all depends upon the application we are talking about.

If we are simply talking about a basic SSO solution, then maybe we will have some patience (I mean no longer than 5 seconds, but this would be considered extremely slow). Now, if the application gets upgraded to like an office setting or warehouse for Time and Attendance and Physical Access entry, we want to be verified as quickly ASAP. And here I am talking like 2-3 seconds. But, if we are now talking about something extremely gargantuan like a national database such as AFIS, we need verification/identification in just one second or less. And as you can see again from the above quote, we are talking of about literally hundreds and hundreds (even thousands and millions) of biometric templates which need to be matched at virtually light speed. And then of course, the metrics then come into play (specifically the FAR, FRR, ATV, and the ERR).

Now, let me point out that in these scenarios I have just described, I am just talking about one type of biometric technology being used. All of this stuff becomes even more needed when we are talking about multimodal biometric solutions-when fingerprint/iris/vein pattern recognition are used. There can be no hiccups in the data connections, or any type of lag time in the processing of the biometric templates.

In the industry, this is known as flushing: “ . . . a specific procedure that periodically pauses the processing of identification requests to optimize internal data structures.” (SOURCE: http://www.findbiometrics.com/industry-news/i/9346/). And even further examples of data management in Biometrics include hardware compatibility, parallel processing, data synchronization, and the list goes on. As mentioned previously, the first image conjured up when thinking about Biometrics is a big lunky device, and Big Brother watching. But there is much more to that obviously, and hopefully with this posting, all of you out there can get a better insight as to just how important data and data management is to our industry.

We all keep talking about standards and best practices for the actual technology itself, but it is also equally important, if not more for the biometrics templates, the supporting algorithms, and how they are all managed.