Impacts Of Memory Virtualization On Real-Time Computing
A Non-invasive Analysis of the Impacts of Memory Virtualization on Real-time Computing
Real-time computing prevalent in industrial control systems. In order to meet the timing requirements, the systems are usually built on specialty components with high determinism. As the capabilities of the hardware advances, it is desirable to integrate multiple systems into a single physical machine to lower the operating and maintenance costs. Virtualization is the well-known technology for consolidating multiple systems by multiplexing various resources among systems. However, the resource sharing of virtualization conflicts with the determinism of real-time computing. It is challenging to find the balance for real-time virtualization because the analysis …show more content…
Wireless sensor networks are probably the most well-known examples. They have been widely deployed in various areas from environmental monitoring, surveillance, to healthcare monitoring and many others [4][5][6]. A wireless sensor network involves data acquisition and transmission from sensor nodes and data analytics in the service platform. All of the stages have different level of real-time requirements. The idea of wireless sensor networks is further extended to Internet of Things. Unlike wireless sensor networks where data are produced by specific types of sensor devices, in Internet of Things, data may come from any devices connected to the Internet and the real-time aspects are more complicated …show more content…
In virtualization, there are several virtual copies of a physical resource, and the access to the resource is usually multiplexed in a time-sharing manner. On the other hand, real-time computing expects deterministic patterns of execution. Resource sharing will increase the degree of uncertainty [26]. Therefore, it is preferred to minimize the usage of shared resources in real-time computing. It is still possible to support real-time computing in a virtualization environment. Theoretically, as long as the system does not exceed the maximum utilization bound, it can guarantee no real-time constraints will be violated. However, it is too difficult to determine the exact bound for an arbitrary virtualization
Real-time computing prevalent in industrial control systems. In order to meet the timing requirements, the systems are usually built on specialty components with high determinism. As the capabilities of the hardware advances, it is desirable to integrate multiple systems into a single physical machine to lower the operating and maintenance costs. Virtualization is the well-known technology for consolidating multiple systems by multiplexing various resources among systems. However, the resource sharing of virtualization conflicts with the determinism of real-time computing. It is challenging to find the balance for real-time virtualization because the analysis …show more content…
Wireless sensor networks are probably the most well-known examples. They have been widely deployed in various areas from environmental monitoring, surveillance, to healthcare monitoring and many others [4][5][6]. A wireless sensor network involves data acquisition and transmission from sensor nodes and data analytics in the service platform. All of the stages have different level of real-time requirements. The idea of wireless sensor networks is further extended to Internet of Things. Unlike wireless sensor networks where data are produced by specific types of sensor devices, in Internet of Things, data may come from any devices connected to the Internet and the real-time aspects are more complicated …show more content…
In virtualization, there are several virtual copies of a physical resource, and the access to the resource is usually multiplexed in a time-sharing manner. On the other hand, real-time computing expects deterministic patterns of execution. Resource sharing will increase the degree of uncertainty [26]. Therefore, it is preferred to minimize the usage of shared resources in real-time computing. It is still possible to support real-time computing in a virtualization environment. Theoretically, as long as the system does not exceed the maximum utilization bound, it can guarantee no real-time constraints will be violated. However, it is too difficult to determine the exact bound for an arbitrary virtualization