IoT Edge Computing Issues and Solutions
1. What are the Issues with IoT Edge Computing?
IoT Edge Computing has become a vital component of modern technology, enabling efficient data processing and real-time decision-making at the edge of networks. However, several critical issues need addressing to harness its potential fully. This article explores the key challenges identified by renowned experts and proposes potential solutions to enhance the capabilities of IoT Edge Computing.
Ju Ren, Yi Pan, Andrzej Goscinski, and Raheem A. Beyah (2018) have highlighted several pressing issues with IoT Edge Computing:
a.?Data Distribution and Management:
Efficiently distributing and managing data storage and computing is crucial for IoT Edge Computing's success. As the volume of data generated by IoT devices grows, finding optimal ways to handle and process this data at the edge becomes a significant challenge.
b.?Collaboration with Cloud Computing:
IoT Edge Computing must collaborate with cloud computing to create more scalable services seamlessly. Integrating both technologies can optimize data processing, reduce latency, and improve system performance.
c.?Security and Privacy Concerns:
Protecting data and ensuring users' privacy is paramount in IoT Edge Computing. As sensitive information is processed and transmitted at the edge, robust security measures must be in place to prevent unauthorized access and data breaches.
2. What are Available Solutions to Address the Challenges?
a.?Leveraging Mobile Devices for Edge Computing:
Betriz Lorenzo et al. demonstrated a robust dynamic edge network architecture that leverages mobile devices to harvest untapped resources and mitigate network congestion. This integrated solution enhances the robustness of the network by considering physical, access, networking, application, and business layers.
b.?Flexible Edge Computing Architecture:
Takuo Suganuma et al. proposed a flexible edge computing (FLEC) architecture that addresses the rigidity of traditional edge-computing-based IoT architectures. FLEC's environment adaptation ability and user orientation enhance the system's agility and responsiveness.
c.?Mobile Edge Computing (MEC):
MEC, a type of IoT Edge Computing, offers a flexible and cost-effective mechanism for mobile devices to offload computation tasks to servers at the network's edge. It reduces the computational burden on devices and improves overall system efficiency.
d.?Cognitive IoT and Prospect Theory:
Infusing cognitive capabilities into IoT devices enables them to mimic human behavioral patterns, making better data-offloading decisions. By applying Prospect Theory, users' risk-based data offloading behavior and decision-making can be better modelled, leading to more intelligent and optimized choices.
e.?Partial Offloading Under Prospect Theory:
In real-world mobile applications, users do not always adopt risk-neutral behavior. Instead, they likely demonstrate different actions under losses or gains regarding their true utility. We use the Prospect Theory to capture the device-centric risk-based decision-making in the MEC environment. According to this behavioral model, individuals make decisions based on risk and uncertainty associated with the payoff of their choices estimated with some probability. Thus, we evaluate the users' actual utility with respect to a reference point. We consider this reference point as the zero point (ground truth) of the users' true utility.
Users' prospect-theoretic utility function is a concave function regarding the user's true perceived utility above the reference point.
Prospective Nash Equilibrium Algorithm:
The best response strategy for a user is decreasing the total amount of offloading data.?
领英推荐
The pseudo-code for the distributed algorithm for convergence to PNE can look like this:?
e. Cybersecurity and IoT Edge Computing
Caso et al.(2023) believe that we need a significant shift in the philosophy of IoT solution design and a holistic convergence of IoT and cybersecurity functionalities to build user confidence in the IoT. When the industry can converge the IoT and cybersecurity, we may reap tremendous rewards. Cybersecurity risk for the IoT encompasses digital security to physical security. Therefore, we must address the complete confidentiality, integrity, and available (CIA) framework. Six key results allow a secure IoT environment: data privacy and access under confidentiality, reliability and compliance under integrity, and uptime and resilience under availability.?
f. Privacy for IoT Edge Computing
Kazi Masum Sadique, Rahim Rahmani & Paul Johannesson (2020) believe that using data anonymisation to protect the identity of users and devices and configuring edge devices to collect only the data that is necessary for their intended purpose can help with ensuring privacy in IoT Edge Computing. We can also use encryption to secure data transmitted between edge devices and the cloud. ?
?g. Combining Edge Computing and Cloud Computing for more scalability
Ken Carroll and Mahesh Chandarmouli (2019) state that making balanced use of edge and cloud computing is one of the crucial points in designing and building enterprise-scale IoT solutions. The combination can reduce latency, increase scalability and enhance access to information. In a cloud-only world where data travels hundreds or thousands of miles and latency is critical, edge computing is important in diminishing latency. We could soon be able to process up to 55 per cent of IoT data near the source, either on the device or via edge computing.
?Rong et al. (2021) propose an edge-cloud collaborative computing platform for building AIoT applications called Sophon Edge. It addresses challenges faced by developers in practice:
·??????Heterogeneity: in a large-scale IoT system, the inherent heterogeneity makes connectivity and coordination processes very difficult.
·??????Accuracy: we need to develop and tune AI algorithms sufficiently to understand and interpret data for more accurate decision-making. We also need to refine models in the highly dynamic nature of the physical world.
The Sophon Edge model introduces a unified device model, called “product”. A “product” is an “abstract class” of IoT devices to declare certain properties (data generated on sensors) or behaviors (available actions of actuators). The product can also declare subscribable events demonstrating some significant changes in a certain IoT device. A user may first define several “products” and then attach each device to one or more “products”. We describe every computation task as a pipeline, consisting of a series of data operation steps. We call each step an operator.?
3.?Future Prospects
As 4G and 5G network usage continue to grow, the demand for IoT Edge Computing is expected to surge. Ultra-low-latency and energy efficiency requirements, driven by the emergence of IoT and Tactile Internet, further fuel the promise of IoT Edge Computing. Combining AI algorithms with Edge Computing is poised to revolutionize industries by efficiently processing high-volume data.
REFERENCES:
Please feel free to subscribe to my LinkedIn to learn more about my work.
Subscribe on LinkedIn https://www.dhirubhai.net/build-relation/newsletter-follow?entityUrn=6891134047146397696
If you have any questions or valuable feedback, please feel free to reach out to me.