Can the cloud go all the way to the edge?

Source – cloudcomputing-news.net

There is lots of conversation in the industry at the moment around the impact that edge computing is going to have on the cloud and specifically if it is going to kill the cloud, or at the very least impact its ability to cope.

With the proliferation of the Internet of Things (IoT), and the predicted future ubiquity of sensors, the cloud and associated technologies in their current state won’t be able to cope with the sheer amount of data being generated by machines – and won’t be able to keep up with the need for speed that these devices and sensors will demand.

To put it into perspective, a driverless car generates 10GB of data every mile from details like GPS, street signs and other surroundings. In order for the vehicle to respond to that data, it needs to process information at an incredible speed, in ‘real’ real-time to ensure that it doesn’t crash and makes it to its destination. To process data at this kind of speed means it can’t rely on that data being sent back to a central cloud in a datacentre, and then transferred back to the device to take action. If the car is coming up to a stop sign, that data needs to be processed in milliseconds, and this processing will need to be done close to the device, not in the cloud. It is said that for this very reason a self-driving car will, in essence, be a mobile data centre with the ability to analyse terabytes of information in almost real time.

In his presentation, Return to the Edge and the End of Cloud Computing, Peter Levine talks about how, in the future, we are going to be collecting ‘the world’s information’. We will have smart sensors on everything; from our ovens to our shoes, to our cars and our keys. These devices will sometimes need to speak to each other in real-time or will need to transfer data instantaneously in order to deliver recommendations about whether your soufflé is sinking, or if you need to change your stride pattern when running. This is already beginning with smart watches and home temperature control devices, but it is easy to see how this industry will continue to boom.

The type of data that IoT devices and sensors produce won’t always be simple text-based information that can be sent back and forth quickly over a network. It will be images and videos that demand lots of processing power, without compromising on speed.

With this move to smart devices in every area of our life, many are predicting the end of the cloud as we know it, and a return to distributed computing where the processing is done closer to the edge. Many of the large online TV and film streaming services are already moving their IT systems to regional hubs to be closer to users in particular areas across the country, but in order to function optimally in this new sensor-driven world, the edge has to come much, much closer.

As technology improves, the ability to monitor more and more physical assets will increase, which will, in turn, increase the amount of data that needs to be processed. For data to be useful in our day to day life, we will want instant insights and recommendations – we will quickly lose the patience of waiting ten seconds for data to be sent to the cloud, processed, analysed and sent back to our devices – much like our impatience when waiting for a website to load. As we become increasingly dependent on these devices, we will need it instantaneously, to provide guidance, recommendations, insights and for the devices themselves to take automatic action.

But what connects everything? Technology. Now, we don’t think that tech – certainly in the sense of data centres or cloud – impacts on our trainers. But in the future, when they’re smart-connected to our Fitbits and phones, trainers will be supported by tech and data. As much as we already talk about our dependence on our phones and technology in everyday life, the world will only become more reliant on tech. From our actions to our travel, to our physical devices, to our cooking and our decision-making.