State of the Edge 2020 highlights the transition we need to make from human to machine speeds in order to enable the Third Act of the Internet. More and more machines are coming online to support the exponential growth of real-time applications and smart devices, from drones to smart home technologies. These applications and devices typically need to communicate with one another and in order to work at their optimal performance level, they require machine speeds.
The original Internet was designed to move at human speed: humans loading web pages, humans reading email, humans watching movies.
“This is why today’s Internet—while fast enough for most humans—appears glacial when machines talk to machines.” State of the Edge 2020
The explosion of the Internet of Things (IoT)
According to IDC’s most recent forecast, there will be 46.1 billion connected IoT devices, or “things”, by 2025, generating 79.4 zettabytes (ZB) of data. As the volume of connected IoT devices grows, the amount of data they generate will also grow. Some of the data will be “small and bursty”, indicating just one metric of a machine’s health. This is in contrast to the very large amounts of data being generated by other “things”, for instance, surveillance cameras using computer vision to analyze crowd behavior.
Machine to machine communication (M2M)
What is machine to machine communication? According to Wikipedia, machine to machine communication is “direct communication between devices using any communications channel, including wired and wireless.”
M2M technology is crucial for developing connections across different aspects of the physical world. It can be used to actuate industrial processes, reduce costs by minimizing maintenance and downtime, and proactively monitor organizational assets to automate business processes and improve customer service.
M2M use cases
M2M was first adopted in manufacturing and industrial settings in order to help remotely control data from equipment. M2M has since found use in a range of sectors, including healthcare, insurance and business. It is also at the foundation of the Internet of Things and used in a range of real world settings, including:
- Industrial instrumentation: remote sensors sharing real time information with application software that uses it to take action - e.g. to share a drop in temperature to adjust an industrial process; to place an order to replenish out-of-stock inventory; to communicate fuel levels in oil drilling machines;
- System monitoring: the use of wireless technology to gain precise data about a system and act as a quality control - e.g. by monitoring a utility meter, the owner can find out if certain components have been tampered with while utility companies can use the information to bill customers the exact amount of resource they have consumed;
- Advertising: the use of wireless networks to update digital billboards, enabling advertisers to display a different message according to the time of day or day of the week, or to respond to a global event, such as a drop in the price of gas;
- Traffic control: sensors being used to monitor variables such as speed and volume of traffic, which is used by devices that control variable traffic information signs and traffic lights.
Latency sensitive vs. latency critical applications
There is an important difference between latency sensitive and latency critical applications.
A latency sensitive application is one in which latency clearly improves performance, but which can still function at higher latencies than optimal, for instance, image processing or bulk data transfer. By contrast, a latency critical application, such as an autonomous vehicle or the controlling of a critical M2M process, is one that will function destructively or fail to function entirely if latency is higher than a predefined threshold.
To illustrate the importance of machine speeds to latency critical applications, State of the Edge cites the example of a robotic drone flying at 60 miles-per-hour, allowing it to traverse the length of a football field in four seconds. Avoiding collisions in a situation like this may require decision-support from an edge server because a delay of 100ms could lead the drone to crash into something only 10 feet away.
The critical imperative for machine speeds and demand for ultra low latency in leisure activities as well, such as VR or multiplayer online gaming, is driving the demand for edge computing and 5G.
The necessity for an edge-enabled Internet
More and more devices creating more and more data, which demand processing with ultra low latency requires a different kind of Internet to the one we have today. Centralized cloud data centers are simply unable to deliver the low latency necessary to keep up with demand from latency sensitive and latency critical applications.
Data can’t move faster than the speed of light; requests to servers in centralized data centers hundreds or thousands of miles away inevitably take tens to hundreds of milliseconds to fulfil. If you’re simply scrolling through Twitter feeds or sending email, the difference can be unnoticeable. However, for a gamer in virtual reality or a surgeon operating remotely, those milliseconds matter.
Using an edge data center for processing of information to reduce latency and jitter on the round-trip can make all the difference. Furthermore, working with an edge compute platform can help businesses significantly reduce latency, deliver consistent performance, and leverage developer workflows to make edge programming a reality today.
At the recent Edge Computing World conference, there was much discussion from analysts, investors and technologists around the need for an edge-enabled Internet.
“In the next few years I think we’ll have a massive switch to the edge. AI is potentially going to dominate all other forms of programming, and AI needs the edge. It’s difficult to build an intelligent system without an edge presence.” Steve Jurvetson, Future Ventures Founder and chip engineer who was an early investor in Tesla and SpaceX