Virtualization technology enables multiple virtual machines (VMs) to share a single physical server. Commercial servers increasingly use the nonuniform memory access (NUMA) architecture due to its scalable memory performance. However, multiple VMs running on a NUMA physical server will cause performance overheads such as remote memory access latency and shared microarchitectural resource contention, which makes the VM performance less efficient and stable. These performance overheads are mainly caused by memory traffic from data-intensive workloads. In this paper, we propose a traffic-aware VM optimization (TAVO) scheme on NUMA systems. Based on the performance monitoring of the data traffic and CPU/memory resource usages in the system, TAVO addresses VM memory access locality and shared resource contention problems via automatic VM initial placement and NUMA-aware VM online scheduling. Our experimental results show that TAVO improves VM performance in terms of benchmark runtime by up to 22.6% compared with the default KVM CFS scheduler. TAVO also achieves a much stable performance with benchmark's average runtime variation under 3%.

Hyperspectral imaging is a non-destructive method for the detection of chilling injury in fruit. However, the limitation of this technique is the lacking of an appropriate working parameters and a feasible discriminating model for chilling on-line sorting. This research was aimed to select the optimal spectral resolution, scanning speed, and classification model for green jujube chilling injury detection based on hyperspectral reflectance imaging. Criminisi algorithm was firstly carried out to reconstruct the specular reflection region in spectral images before deriving mean spectra, and thus the optimal wavelengths were selected by random frog. Results showed that the Criminisi algorithm presented a desirable ability of spectral image inpainting. The linear discriminant analysis (LDA) achieved overall accuracies of 98.3% and 93.3% for two-class and three-class classification, respectively, at the speed of 20 mm/s with the spectral resolution of 5.03 nm based on selected spectral features. The results demonstrated that 20 mm/s with the spectral resolution of 5.03 nm was more feasible for the detection of green jujube chilling injury in hyperspectral imaging system due to a higher scanning efficiency, but a less data size.