Getting an ‘edge’ on the IoT will require using local storage to tackle limitations of the cloud

By Charlene Wan and Vivek Tyagi

Cloud is all about scalable, secure and long term data retention; however, it is not the only place where storage is critical. Local storage is important for devices at the edge to capture data and provide real time analytics support.

Internet of Things is undergoing an explosion and every day it becomes more and more critical to deal with the requirements of this expansion. We’ve all heard numbers being thrown about, 50 billion connected devices expected by 2020, which would generate tons of zettabytes of data that needs to be stored and managed. Although, many would assume that all this data can be sent directly to the cloud, the realities of storage solutions for IoT are far more complicated.

Consider the volume of data created just in transportation alone. Every autonomous car on the road by 2020 is expected to generate 2 petabytes (PB) of data each year and a single airplane will generate 40TB of data daily.

Now pushing all this data onto the cloud would be difficult. Even if 5G networks become a reality, there might not be enough bandwidth for real time data transfer. This has led connected cars and planes to make use of on-board local storage for capturing data and creating a cache, till the time it can be moved to the cloud via high speed networks.

Let’s take another example - consider manufacturing facilities, the data generated per day is projected to be 1PB. Smart buildings generate another 250GB of data per day and smart utility metering systems generate more than 3 exabytes of data. In theory, these systems should be able to transfer all this data onto the cloud in real time when connected to high speed network.

However, is cloud the best place for all this data?

So, is the cloud really the best place to keep all that data? The not so simple answer is, not always.

Although the cloud is considered to be elastic and ubiquitous; and has analytical tools to help discover valuable information from large pools of data. There are still some limitations to it, real time analytics cannot be performed on cloud based data, and by the time the data is transferred to the cloud it becomes incompetent. Most importantly in case of critical conditions, like a sensor detecting out of bounds condition on a shop floor device, real time information becomes pivotal.

There are a number of situations where an instant response is necessary, this is where real time analytics runs near the edge of the cloud or the “fog”, these could work as gateways, aggregation systems or more. Through this, automobiles and planes can now be equipped to respond in real time to mechanical, environmental or situational alerts.

Even though IoT has progressed, it is still difficult to ascertain which data is important.

Take for instance a traffic camera on a busy street. There would be hundreds if not thousands of images that would get captured every day. While most of these images may not be useful at all, there would be certain frames that may be critical to understand an accident or traffic violation.

Advanced data intelligence and pattern recognition tools have become more of a necessity for all areas of IoT. These will help in getting an in depth understanding of the inherent value of the captured data, be it in the cloud or at the edge. When talking about edge devices, whether a sensor or a smartphone or a gateway, storage that is fast, reliable, and secure is critical. A high ratio of capacity to physical footprint, along with the ability to withstand harsh weather conditions are also important considerations as edge devices live in a variety of environments.

Charlene Wan is the Marketing Programs Director at SanDisk and Vivek Tyagi is the Director for India business development, SanDisk brand Commercial sales and Support at Western Digital Corporation.

Updated Date: Jul 17, 2017 14:36 PM