Improvement of load balancing mechanism in multi-core architectures over high availability technique
Arsen Kurti, Igli Tafa, Aldi Disha · AIP conference proceedings · 2015
Nowadays Multi-core CPUs have become a trend in personal computers, smart phones and tablet market. Due to high performance of these architectures developers like Google Play Store and Apple App Store have been able to adopt more attractive interfaces and powerful applications. Complexity and size of these applications is getting bigger and multitasking is always used in nowadays computing. But these huge development in software and hardware needs to have good user feedback. To improve the user experience needed to push into the limits of multi-core systems. In this paper will be described how to improve the real time performance by reducing the user wait time, power consumption, memory usage of multi-core systems that use different run queue for different core scheduling technique over private cloud. Traditional way it is not take in the consideration in order to equal the balancing cores but to minimize the cost in time, memory and power consumption of task migrations. This will be possible by creating a importance level of performing task migrations to balance core usage and designing a algorithm to share the tasks into different core queues to create a balanced core usage. This new technique will improve the real time characteristics such as waiting time, cache efficiency and power consumption.