Edge Artificial Intelligence (Edge AI) allows developers to build smart applications that can make decisions using local data on their devices. This means there is no need to wait for communication with a central server; developers can protect users' private information; and developers don't have to rely on expensive cloud services. However, it can be difficult to deploy AI models on devices with limited capabilities (such as small screens, low battery life, etc.) due to the trade-offs between inference speed, energy usage, memory constraints and accuracy of the predictions.

The introduction states that the use of shared cloud infrastructures is now fundamental to contemporary computing because it provides organizations with the ability to scale up their computing power, lower operational costs, and run many different types of applications. On the other hand, the fact that there are multiple users utilizing the same computing resources creates issues such as misuse of resources, poor allocation of resources, illegal usage of resources, and unusual behavior by the workloads. Many traditional monitoring systems use set limits and rules to determine what is normal or abnormal; however, these systems can fail to detect complex misuse patterns or provide an explanation as to why an anomaly was found.

Latency sensitive applications demand fast processing and timely response to ensure correct, reliable and efficient operation of the application. Autonomous Vehicles, Industrial Automation, Health Monitoring Systems, Smart Cities, Interactive Systems etc. usually require strict response times. Traditional cloud computing offers scalable processing power and central control; however, sending data to remote servers introduces delay in transmission (network delay), congestion, and communication overhead. Edge Computing reduces delays by processing information near the source of data, however edge devices are restricted in terms of computing power, memory, and power. This paper proposes a light-weight decision support system to help determine whether to process on the Edge or Cloud based on the requirements of the application and the operational environment.

Cloud computing platforms are now a vital tool used by businesses to host up-to-date applications; process large amounts of data; and provide users with scalable digital services. Cloud workload fluctuates often due to varying user demand; application behavior; scheduled activities; and unplanned traffic surges. Predicting accurately when a short-term increase or decrease in cloud workload will occur can assist cloud service providers to allocate resources effectively; lower operating costs; and ensure application performance.

Containerized applications are now a core component of today's cloud native infrastructure due to their benefits of portability, scalability, flexibility in deployment and efficient use of resources. As a result, there has been an explosion of container technology; however, this growth has also led to many security concerns. One of the main areas where these concerns lie is in regard to too many permissions being given to services and unauthorized access to system resources. Many times applications will have more permissions than they actually need to run properly. This creates greater opportunities for potential damage caused by compromised services, malicious actions or privilege escalation attacks.