Computing on the Edge
Thomas Friedman's 2005 book titled The World is Flat discussed how globalization and technology flatten barriers to entry, making the world more competitive. From 2005 to 2021, there is no shortage of examples to highlight our interconnected, interdependent, and intertwined global economy. However, the world is rounding out again as countries seek to limit how globalized its citizens' data become by passing comprehensive data protection regulations.
Businesses rely on the borderless flow of personal data, and the ever-changing privacy compliance landscape highlights how changes in data flows impact business as usual. As a business dealing with personal data, you either find a way to work within the system or you find a way to avoid the system altogether. Therefore, businesses are tasked with data minimization and localization or finding ways to avoid collecting and using personal data altogether, which, in most cases, is not a sensible solution.
Cloud computing environments significantly flattened the world by providing, seeming unlimited, processing power to multi-national corporations and small businesses in most jurisdictions. The large shift to cloud computing also contributed to the rounding of the world as these environments were impacted by the aforementioned data protection regulations.
Cloud computing made data localization difficult because if we are physically located in Santiago, Chile, then our data is stored in data centers located in the Sao Paulo, Brazil Region; however, our data is routed to the main data center through an Edge Location, which for this example is located in Santiago. This means that the data is stored in Brazil and, if Chilean data privacy laws prohibit certain kinds of data from being exported, there is not an easy way to use a cloud environment given these localization issues. Edge computing is the solution that is helping re-flatten the world, our current model of cloud computing, and solve part of the data protection regulation puzzle facing businesses.
Edge computing is the solution that is helping re-flatten the world, our current model of cloud computing, and solve part of the data protection regulation puzzle facing businesses.
Edge computing will grow in importance as connected devices permeate the market. The Institute of Electrical and Electronics Engineers listed as one of its top technology trends Artificial Intelligence at the Edge, stating "ubiquitous connectivity such as 5G and intelligent sensors such as the Internet of Things (IoT), ML applications will rapidly move to the 'edge.'" Edge Computing will play a crucial role in the connection of a IoT devices such as self-driving cars, industrial wearable devices, and other autonomous robots and drones. Specifically, for self-driving vehicles, Edge computing is essential for processing sensory data given the near instantaneous processing times due to not having to connect to the centralized database. Therefore, the next generation of technology and data analytics relies on the rollout and improvement of Edge computing.
Additionally, Edge computing, with the implementation of certain machine learning tactics, seeks to bolster privacy and work within the growing international regulatory environment. Swarm Learning is a decentralized machine learning process that seeks to solve the data centralization problem posed by cloud computing. A journal article titled Swarm Learning for decentralized and confidential clinical machine learning states Swarm Learning "dispenses with a dedicated server, shares the parameters via the Swarm network and builds the models independently on private data at the individual sites (short ‘nodes’ called Swarm edge nodes)." The article continues that Swarm Learning "provides confidentiality-preserving machine learning by design and can inherit new developments in differential privacy algorithms, functional encryption, or encrypted transfer learning approaches."
In other words, let us say there are four labs in four different countries that want to collaborate in building a model to help predict tumor detection. We assume health privacy laws prevent the labs from uploading all of this sensitive data into the cloud or exporting the data to other countries. In an overly simplistic explanation, Swarm Learning allows each lab to process the data within its geographic boundaries and then the local machine learning model shares parameters learned from its local data set without ever sharing unique data point. Therefore, the models are constantly being updated, refined, and trained based off of the insights of the other labs. Swarm Learning is touted as preserving privacy while allowing for fair and transparent processing of highly regulated data sets.
There is always a downside. Edge computing may increase cybersecurity risks simply because of the sheer number of devices that allow for hackers to attack. However, Edge data centers that are properly architected could be used as a tool for cyber resilience by operating isolation from the core. Essentially the idea is that while an attack may shut down a specific edge location or cluster therein, the entire network would be protected prohibiting widespread contagion. The Institute of Electrical and Electronics Engineers in A Survey on Security and Privacy Issues in Edge Computing-Assisted Internet of Things states, "security and privacy objectives can be met by developing different protection mechanisms for authentication, access control, data transmission, storage, and computation." Therefore, as Edge computing expands, businesses will need to stay on top of best practices for cybersecurity measures given Edge computing opens new vulnerabilities while also providing new solutions.
In words of Thomas Friedman, "big breakthroughs happen when what is suddenly possible meets what is desperately needed." Edge computing will be essential as IoT regulations are rolled out, data privacy regimes internationally become more comprehensive and provide greater protections, and as cybersecurity threats continues to be an industrywide concern. It is time to start asking how your business's data may be impacted by data localization requirements and how Edge computing can address these challenges.