Edge Computing and the Internet of Things
Data and the internet make up an increasingly important part of our lives, as more of the world comes online and technology continues to become more readily available and capable. In the last year alone, about 38.5 zettabytes of data was created, which converts to about 38,500,000,000,000 gigabytes of data in one year. To put that in perspective, if each Gigabyte in a zettabyte were a brick, 258 Great Walls of China, made of 3,873,000,000 bricks each could be built (Cisco). With each passing year, data is produced at exponentially higher rates, with more and more of the population coming online as well as more and more connection and internet based products. With the ever increasing demand and usage of data by consumers around the world, especially now in a mostly virtual and online workspace, it does raise the question on how data processing and internet connectivity can be improved.
Defining the Internet of Things
This is where edge computing and the Internet of Things (IoT) comes in. The Internet of Things refers to the network of connections between physical data receptors, or “things”, which contain things like sensors, monitors, and other technology for recording, connecting, and exchanging data over the internet. For example, smart devices like Amazon Alexa and smart watches like the Apple watch all record and share data that they collect, whether it be to your phone, laptop, or company servers. With the production of cheaper and smaller computer chips, almost anything can be turned into part of the IoT, enabling massive amounts of data being communicated without the involvement of people.
The size of the IoT continues to grow with each passing day, as more and more “things” are added to commercial and residential spaces, especially with the rising popularity of smart devices. Tech analytics company IDC predicted there would be 41.6 billion devices connected to the Internet of Things by 2025, with these devices most likely being from the industrial and enterprise side, as production and vehicle equipment have consistently made up a massive part of the IoT. However, everyday people and consumers have also made up a significant part of the IoT, though at a much lower rate than massive global corporations.
Advantages of the IoT
There are many pros and cons to the IoT and its ever-expanding reserve of data receptors and transmitters. Primary benefits mostly lean towards corporate enterprises, as the collection and transmission of data in large quantities would be critical to increasing speed and efficiency in a workspace. Companies like Facebook, Amazon, and Apple have competed in this space for a long time, producing many products collecting, connecting, and receiving data. These products often make life a lot simpler for consumers and businesses alike.
Companies often integrate sensors and receptors into products to monitor their efficiency and function, allowing for a real-time monitor on product functions. Based on this data, products can then be improved or updated to better suit needs or drawbacks. For everyday people, the Internet of Things offers many conveniences through smart devices. Smart homes offer automation of many day to day functions like air conditioning and lighting, while smart devices like smart fridges and smart kitchens allow ease of access with internet connectivity.
Smart homes and devices rise in popularity significantly with each passing year, integrating more and more into tech culture and developments. For example, Apple AirTags are a good example of a modern “thing”, featuring a comprehensive location service, audio location service, and directional and gyroscope based locating features. With all this data, the AirTag can make life significantly easier should anybody want to find a lost item, whether it be keys, a wallet, or any other personal belonging. Additionally, statistics have shown that nearly 70% of Americans already own a smart appliance, while the American smart appliance market reached 23 billion dollars in 2020 (Techjury).
With the massive and increasing popularity of smart devices in America, there’s no doubt that the IoT will only expand as time goes on and more and more technology becomes available to the public. The expanse of the IoT indicates the automated and regulated flow of more and more data, but does not solve the problems of speed and agility.
The IoT and 5G
The Internet of Things revolves pretty heavily on connectivity methods, whether it be Wi-Fi, Bluetooth, Ethernet, or even LTE. The development of faster and more efficient, even interchangeable options for connectivity, was a key point of improvement in IoT interactions. The IoT’s development was tightly connected to the development of Wi-Fi and new connectivity methods.
Thus, with the expansion of 5G networks and connections, IoT projects and data collections can be greatly expanded. By allowing more and more devices to exist in a concentrated area thanks to improved and secure network connections, the number of sensors or “things” can grow exponentially. With the introduction of edge computing to these large networks, necessary and unnecessary data can be filtered and distributed quickly, effectively, and locally, leading to many benefits for processing and analysis of data.
Edge Computing
As a solution to the desire for faster connectivity and lower latency, edge computing and 5G offered many benefits that complement the increased data volume that the IoT provided. Edge computing was developed primarily after the exponential growth of information services and IoT devices.
Edge computing can be defined as a distributed computing model where computation and data storage are brought closer to the “edge”, where the data is being collected and used. In typical cloud computing architectures, information is received and gathered at large centralized data centers, which were often thousands of miles away from their target devices. However, by bringing the computation and storage aspect closer to the “edge” or devices, latency and connectivity issues can be significantly reduced. Edge computing hardware and software can also considerably solve latency issues by being local sources for data to be received or processed. Edge-computing hardware is also able to process data from a set number of devices, determining the relevant data, and then send only the necessary or desired data to the cloud, which would help significantly in the reduction of bandwidth consumption.
Pros and Cons of Edge Computing
Edge computing can thus lift a lot of dependency off of cloud computation, offering faster and more effective response times, as well as reduced bandwidth usage. As an alternative approach to data networking, the abundant processing power of the IoT devices and edge data centers can be used to locally and quickly process data. Edge computing thus completely contrasts cloud computing and its architecture, which allows for large scale data reception and processing through huge storage and processing hardware and software. Data analysis in cloud computing, however, is one primary limitation of edge computing, as edge computing is limited to the local information accumulated in each receptor or “thing”.
Though edge computing will likely not completely replace or surpass cloud computing, it offers an alternative and more specialized usage of the IoT and architectures of local processing and storage. Through this alternative approach, companies can take large strides towards processing and storing data quicker, enabling higher efficiency in regards to real-time data and applications that rely on this data. With edge computing models, algorithms that would have to run on the processing and computing of a central cloud server thousands of miles away can be run locally on an edge server or even on the device itself. By shifting to an edge based approach, lower data transfer costs and the ability to have more reliable real-time data can be obtained, but the edge computing approach is not always the best. Though the reduction of transport and storage amounts in traditional methods would be very desirable to large scale companies dealing with large flows of data, cloud based architectures are still extremely effective in dealing with even larger amounts of data and processing it all together in “big picture” functions and algorithms, or at least much more so than locally limited edge computing.
The growth of edge computing isn’t necessarily at the cost of cloud computing either, as both edge and cloud computing are quite compatible. Integrating edge features with the already existing and predominant cloud infrastructure would lead to a “best of both worlds” situation, with companies being able to allocate more towards edge or cloud computing based on their needs. By developing more effective and faster data processing and reception, not only does the company advance, but so does technology and other features for the consumer down the line.
Applications of Edge Computing and the IoT
AI modules like the Jetson Xavier NX module (above) by NVIDIA were developed to take advantage of edge computing. Though a very small and versatile module, the Xavier NX can still run fairly complex AI algorithms. These algorithms typically need massive amounts of processing power, leading to the popularity of hosting and running them off of cloud data centers. However, with the continued research and development of smaller and more advanced AI chipsets, edge computation can only improve, allowing for faster and better responses in instant computing. The production and usage of more of these AI capable chips bring the possibility of the IoT shifting to an AIoT, allowing for improvements in robotics hardware and software, machine vision, and data analysis.
Companies like AWS, Google Could Platform, and Microsoft Azure are all developing and starting use of edge computing capabilities, understanding the potential successes of an effective integration of edge and cloud computing. This collaboration and mix of the two computation approaches would be beneficial to the development of data exchange and processing, especially in regards to the ever-expanding Internet of Things.
Additionally, integrating the IoT and different approaches to data processing like edge and cloud computing could offer the solution to many future projects. For example, the development of autonomous vehicles would be directly connected to interactions of real-time data and devices. While the infrastructure and technology for completely automated vehicle systems, smaller demonstrations of the concept have proved its viability in the real world. Automated truck platooning, demonstrated in projects by Volvo and Peloton, has been tested and used in real life scenarios. With the improvement of connectivity and local interactions of data receptors and processors, these truck platoons can become more reliable and useful in real driving situations.
Potential Setbacks and Obstacles
With the advance of technology comes the rise of new solutions and new problems, especially seen with the development of edge computing. Security is a main problem with edge computing, as local devices may not be as secure as the traditional centralized cloud facilities. Additionally, with the expansion of the IoT, the amount of devices, each with their own security measures, would require attention and care to solve potential security issues in the devices and their security.
After the initial developments of edge computing technology, security became a primary setback and problem in the future of edge computing, resulting in significant time taken to establish secure connections across thousands or millions of devices. However, as more solutions are achieved towards this goal of safety, edge and cloud computing can be streamlined and used more and more reliably, resulting in the overall increase in speed and analysis of the thousands of gigabytes of data created a year.
Concluding Thoughts
With the advent of the Internet of Things and edge computing comes the possibility for the development of many possible applications. For example, autonomous machinery and vehicles could eventually be attained with the use of real-time information and data processing and analysis, established primarily by interactions of the IoT and local “edge” processing. To reach these goals, however, issues like hardware problems in electricity and connection, as well as the frontier development of solutions to security problems, would have to be resolved. With their resolution, however, comes the promising future of more data and automation involved in making the human life better and easier.
