Ready for the disruption from edge computing?
By Scott Amyx | 7 minute read | August 29, 2016
Technology evolves incredibly quickly—it seems that just a few years ago, companies were struggling with the idea of moving part of their operations to the cloud. Now, the cloud is seen as a primary driver of disruption in business, allowing even the smallest, leanest company to have access to incredible processing power and data insights. However, if there’s any lesson that can be derived from the move to integrate cloud technologies, it’s that business will never be the same. Companies should prepare for another renaissance motivated by edge computing. Like the move to the cloud, edge computing is creating vast new opportunities and dismantling entire approaches to business. Decentralization coupled with intelligent computing will have an impact on every industry, including insurance (what do you charge for driverless cars?), banking (what if people use crowdfunding instead of loans?), and mining (robots instead of people?), as well as governments (secure voting, anyone?) and organizations (need to organize without interference?). What is driving this change and the possible impact are important ideas that need to be explored in order to develop strategies and business models that can compete in a high-technology marketplace.
Living on the edge
Edge computing (sometimes referred to as fog computing) involves pushing data storage and processing to the “edges”, or remote parts, of a network. It allows for actions to be taken on-site, at the point of processing. Right now data is focused in the hands of the few. Networks are largely centralized, with data moving from the edges of a network to a center, often a cloud server. Companies and organizations collect and store data – even data that may have little value to them. As consumers have become more tech-savvy, they have begun to demand greater data privacy and control (European regulators have moved swiftly to address concerns through General Data Protection Regulation). Decentralization can offer consumers what they want: control over their own private information, a safe way to manage that information, and a secure way to interact with others. A number of companies have attempted to create private storage areas for consumers or develop programs to encrypt data, but few can really provide the safety and security offered by a decentralized network. Decentralization is a significant transfer of power. Not only will it improve privacy and security, but it will also cause a significant disruption in industries and business models.
Architects of disruption in real-time
Right now, there are two key drivers for the decentralization that comes from edge computing. The first is the need for real-time processing. Currently, there are a number of issues with processing times. They’re fast, but not fast enough, especially if data needs to be sent back to a cloud for analysis. There are not many Internet of Things (IoT) sensors that can manage this type of fast processing on-site, but that is about to change: MIT’s Eyeriss and Intel’s Curie are just two examples of the incredible computing power that will soon be available in the palm of your hand or even smaller. Advances in technology have yielded novel, enhanced ways of transmitting data, not only allowing for the rapid exchange of information between a server and a device, but also between devices. Twenty years ago, driverless cars were science fiction, but now Google’s autonomous vehicle is navigating roads and using LIDAR to help the vehicle avoid crashes and stay safe on the road. MIT’s jumping cheetah seems a little like an engineering student’s dream toy until you realize that the machine is making real-time decisions—quickly and autonomously, without human intervention—to navigate obstacles. IoT is a part of the mix, too: sensors that are attached to just about everything can create a vast, scalable network of incredible power and size that is completely decentralized.
Rise of the networked machines
The second driver of edge computing is related to AI/machine learning and the creation of a neural network (a computer networking model based on the biological network found in the human body). Mesh networks are a considerably democratizing force in an age of centralization. Typically associated with political protests in countries where speaking out is extremely dangerous, mesh networks allow for a user to move their transmission from one device to another, completely eliminating the need for a cell tower. There are, of course, other uses for mesh networks, as they offer cost and latency advantages to users. These networks will be increasingly used by artificial intelligence. The snowballing number of devices, sensors, and machines that need to make real-time decisions will further motivate the decentralization of networks. Autonomous agents are vital to a distributed mesh network since there may be areas where there is no way to access the Internet or not enough time to access a remote server to deliver a key decision.
Currently, the cloud is providing the power of analytics to companies of every size, and companies utilizing analytics are enjoying a significant advantage in the marketplace. However, the movement towards local, real-time analytics is already occurring. For example, MIT’s Eyeriss uses no Wi-Fi and requires no data transmission—it can operate without the Internet. Intel’s Curie is as small as a button but has a tremendous amount of processing power. These devices need to be intelligent in order to be able to aggregate data and deliver solutions in real-time. Living algorithms can work locally, thereby offering a higher level of performance and complex computations only previously available with the cloud.
Blockchain is not just for Bitcoin. Although Bitcoin is the largest supercomputing network in the world, it has managed to address the key issue behind any transaction: trust. This decentralized system has incredible potential. Blockchain operates through a system that utilizes complex algorithms to help prevent fraud and theft; while the system is far from perfect (for example, there are issues with processing times), its potential reaches far beyond managing currency transactions. For example, voting can be made simple, secure, and private through the utilization of decentralized technologies (Agora currently offers citizens a voting method that utilizes open source while the GES/Blockchain Technologies partnership is advancing its own form of digital democracy). Domain name registration can also benefit from decentralization (for example, Namecoin and its .bit domains) because the top-level domain will be owned by no one. From a privacy point-of-view, the expansion of private, safe storage space will be a strong benefit of the decentralized network: the blocks in the blockchain are eminently versatile. Finally, Dapps (decentralized applications that make use of blockchain technology) will have a broad effect on the automation of the network. Of course, these networks will need to be incredibly fast and responsive, so the need for low latency systems is vital to the success of edge computing. Right now, the Internet seems fast (well, mostly). However, latency does matter, and it has a significant impact on real-world decisions. The development and enhancement of networking protocols is lagging and the current alphabet soup available makes it difficult to imagine a worldwide network capable of fast response times across a heterogeneous standards and protocols. However, the move towards open source solutions could have a strong impact on this area.
The age of redefinition
We are not only in the midst of redefining how industries will manage the disruption from decentralized networks and the rise in autonomous systems, but we are also in the heart of the complete overhaul of the Internet as we currently understand it. The Internet is a global computer network consisting of interconnected networks using standardized communication protocols. These networks include data centers, cloud servers, academic research computing labs, and individual desktop machines. That network topology is about to change to a wireless sensor network that is peer-to-peer (P2P). This revolution in topology will affect an array of industries, including telecommunications. Traditional telecom carriers and broadband providers (who currently enjoy an oligopoly, and in some countries, monopolies) will need to re-evaluate how they approach their business. Mesh networks will allow for the bypass of a telecom’s gateways and cell towers. Just over a decade ago, mobile phones were the domain of the privileged. Now, everybody has a smartphone—not to make calls or take photos, but to make payments, arrange appointments, check on inventories, examine statistics… people use their phones for everything. With the rise of AI, these devices can be used for any number of activities, since the increased processing power and network capabilities once only enjoyed by the few will be available for the many – without prejudice.
Edge computing is forcing change, both in the way information is sent or stored and in the way people interact. Business models will need to change in order to adapt to this high-technology, decentralized marketplace. Basically, the era of robots and autonomous smart things is upon us. Just look at Viv. The complete transition to the edge will take time, but the future of edge computing is not that far off. Edge computing’s primary gift is that of power: it re-distributes it.
Learn more about how IBM and Cisco are bringing the cognitive power of IBM Watson to devices at the edge of the network. Visit http://ibm.biz/totheedge.
Scott Amyx is a guest blogger from the IBM IoT thought leadership program.