Maintaining Data Value Through Edge Analytics
Exponentially more data is being produced now than only a few years ago. Much of this data is being created as the result of the Internet of Things (IoT), and quite a lot of it is also created by mobile users. In just three hours during the Super Bowl, for example, mobile device usage was responsible for the creation of more than 22 terabytes of data.
The speed at which data is created and replaced by new data has forced analysts to the realization that data is perishable — it has a shelf-life after which it is no longer useful or valuable. The sheer volume of this data has also necessitated changes in how and where data is stored and analyzed. Twenty-two terabytes is a lot of data, but it isn’t any use at all if it can’t be analyzed, and by the time that much data could be transmitted to a data center, the insights to be gained from it would be out of date — effectively useless.
Edge Analytics to the Rescue
The IoT has created new sources of data that have, in turn, resulted in vast oceans of potentially useful data — data that is, unfortunately, perishable. Rather than surrender to the fact that getting that data to a data center would take so long that the data could no longer be acted upon, however, smart businesses are making use of something called edge analytics.
In addition to the large, central data centers they have always used, businesses are starting to add smaller data processing hubs nearer the edges of their networks, allowing smaller sets of perishable, local data to be analyzed as quickly as possible.
One of the key factors that makes edge analytics a useful technology is the speed and accuracy of modern data tier sorting.
For a long time, companies have sorted their data in terms of how important it is, how sensitive it is, or how frequently it is accessed. Adding time sensitivity to that list helps to ensure that perishable data is analyzed and utilized before it can become meaningless.
Analysis performed on IoT data is most often predictive analysis used to anticipate a future condition and provide recommendations for best making use of those conditions. Due to the nature of most IoT devices, in particular their often rapidly changing environments, the future they are meant to predict is often only moments ahead of the present.
In order to capitalize on the availability of this type of data — to target a customer with sales information as he moves from one part of the store to another, for example, or to analyze data from home health monitors so as to spot worrisome conditions before they happen — it is vital that the data be analyzed as close as possible to where it was generated and collected. Therein lies the value of edge analytics.