Is cloud decentralisation inevitable?

The Internet of Things (IoT) is growing rapidly, disrupting every industry and many aspects of our daily lives. Devices and appliances that were not connected (from fridges and thermostats to cars and watches) are now equipped with sensors, processors, and connectivity.

It is obvious that IoT will continue to grow exponentially over the coming years. Gartner predicts that, in the not so distant future, almost all things will eventually be connected to the internet: by 2020, IoT technology will be in 95% of electronics for new product designs.

IDC also forecasts an astonishing growth for IoT as enterprises become increasingly eager to take advantage of its benefits. According to a recent IDC study, worldwide spending on IoT will reach $772.5 billion in 2018 and will top $1.1 trillion in 2021.

>See also: Why 2018 is the year of digital decentralisation

In addition to the spectacular growth of connected devices, another trend is leading to a data deluge: data generated by users and machines at the edge of the network. Mobile internet and social media have empowered people to become producers of data. Every minute, more than 400 hours of video is uploaded to YouTube, over 243,000 pictures are posted on Facebook, and more than 29 million messages are sent via WhatsApp.

This is causing a shift in the global data traffic pattern: from consumer devices that mostly consume content (hosted on the cloud) and generate downlink traffic (from the cloud to the device) to powerful devices that produce content and generate significant uplink traffic.

This shift is stressing both the network and central cloud infrastructure. The same holds true for machines. According to Intel’s CEO, an autonomous vehicle generates 4 TB of data per day, the equivalent of the amount of data traffic generated by 3,000 people in one day.

Major challenges to cloud computing

The explosion of smart devices and data generated at the edge results in significantly more demand on the computing infrastructure than the current centralised cloud infrastructure can handle.

The current network and computing infrastructure cannot possibly manage to send all the data back to the central cloud to process. We also need orders of magnitude more resources in central cloud to process and analyse the data on servers in remote data centres. Even then, in most cases, we would have to transport the results back to edge devices for actionable reporting.

>See also: Why identity management should be decentralised

We are amid a data explosion at the edge but the tsunami is yet to come. The Internet of Things is growing at a fast pace and is already bumping against networks’ physical limits.

Historically, computing power has roughly doubled every 18 months (or a hundred times every decade) whereas network bandwidth has grown about 50 times per decade. In other words, even if the number of devices remains constant (which clearly won’t), the communication network will be the bottleneck for the growth of IoT.

Moreover, even if network capacity was miraculously increased to cope with all that data, laws of physics will inhibit remote processing of data in the central cloud due to large latencies in the long-haul transmission of data. Reducing latency even by a few milliseconds will be of uttermost importance for many IoT applications. For instance, in an autonomous car, a few milliseconds could be the difference between life and death.

We should decentralise the cloud today

Thankfully, there is a way to turn tens of billions of devices from a challenge to an opportunity. We need a paradigm shift to unleash the power of computing devices at the edge.

There are vast amounts of unused computing resources at the edge that sit idle most of the time. For instance, the collective computing power of 50 million Sony PlayStation 4s in the market today is roughly four times that of Amazon Web Services 27 data centres around the world. There are billions of PCs, tablets, set-top-boxes, smart routers, connected appliances, and smartphones which collectively have millions of times the computing resources of central cloud.

>See also: Will innovation lie in the cloud, on the edge or elsewhere?

If we can harvest the unused resources on these computing devices, we can create an edge cloud that is millions of times larger than the entire central cloud resources today. This is the only scalable solution to address the skyrocketing growth of data.

In other words, we need to evolve to edge cloud computing, a decentralised architecture where boundaries between servers and clients are broken and resources on central cloud and the edge are combined to form an expanded cloud fabric. Edge devices can process data locally, communicate with other devices directly and share resources with other edge nodes to unburden central cloud computing resources.

Edge cloud is the natural and inevitable evolution of cloud computing. The architecture is more efficient and scalable and can deliver lower latency. In many cases, data can be immediately analysed and put into action overcoming bandwidth limitations and latency constraints on the network. It also entails relevant social and economic implications. Decentralised cloud is more private in nature since it minimises central trust entities and is more cost-efficient since it leverages unused computing resources at the edge.

Will edge cloud replace central cloud?

Make no mistake: edge cloud computing is not a replacement for central cloud. Some applications are better suited to computing resources in data centres. However, central cloud (servers in data centres) should not be the bottleneck. They should be used only when there is merit. Servers in data centres should work along with edge devices to build a distributed edge cloud architecture.

>See also: The evolution of cloud: the ‘intergalactic’ infrastructure of the future

Delegating some of the cloud’s processing and storage out to edge devices (near to where the data is acted upon) is essential for many applications that require low latencies and hyper-localisation.

The increasing enterprise adoption of IoT will make it even more necessary to decentralise the cloud. According to Gartner, around 10% of enterprise-generated data is created and processed outside a traditional centralised data centre or cloud. This figure will reach 50% by 2022. The potential impact of edge cloud as an enabler of the hyper-connected world is immense, bringing massive opportunities to enterprises and benefits to users.

 

Siavash Alamouti’s – CEO of mimik – space-time block code (the Alamouti code) is included in over 10 billion devices

Avatar photo

Nick Ismail

Nick Ismail is a former editor for Information Age (from 2018 to 2022) before moving on to become Global Head of Brand Journalism at HCLTech. He has a particular interest in smart technologies, AI and...