HMvisor:dynamic hybrid memory management for virtual machines

来源 :中国科学:信息科学(英文版) | 被引量 : 0次 | 上传用户:liuwuguigui
下载到本地 , 更方便阅读
声明 : 本文档内容版权归属内容提供方 , 如果您对本文有版权争议 , 可与客服联系进行内容授权或下架
论文部分内容阅读
Emerging non-volatile memory (NVM) technologies promise high density,low cost and dynamic random access memory (DRAM)-like performance,at the expense of limited write endurance and high write energy consumption.It is more practical to use NVM combining with the traditional DRAM.However,the hybrid memory management such as page migration becomes more challenging in a virtualization environ-ment because virtual machines (VMs) are unaware of the memory heterogeneity.In this paper,we propose HMvisor,a hypervisor and VM coordinated hybrid memory management mechanism to better utilize DRAM and NVM resources.HMvisor exposes the memory heterogeneity to VMs by mapping virtual NUMA nodes to different physical NUMA nodes.We propose a lightweight and efficient page migration mechanism by decoupling page hotness tracking from page migration.HMvisor performs those operations in the hypervisor and VMs separately,without disrupting the execution of VMs.We also propose a memory resource trading policy to adjust the capacity of DRAM and NVM for each VM,with the monetary cost unchanged.We implement our prototype system based on QEMU/KVM and evaluate it with several benchmarks.Experi-mental results show that HMvisor can reduce 50% of write traffic to NVM with less than 5% performance overhead.Moreover,the hybrid memory adjustment scheme in HMvisor can significantly improve application performance by up to 30×.
其他文献
It is conducive to the sustainable development of human beings in karst regions to research the mechanism of karst rocky desertification(KRD)expansion.Whether the large-scale KRD in southwestern China is caused by climate change or human activities is sti
This paper presents a scheme for simultaneous fault estimation and fault-tolerant control of linear discrete time-varying stochastic systems.An observer is proposed to estimate the system state and the fault simultaneously.The estimation errors of both th
This paper presents an adaptive state-feedback strategy for state-constrained stochastic high-order nonlinear systems.By adding a power integrator and adaptive backstepping techniques,a new adaptive controller is constructed without imposing feasibility c
Real-time automata (RTAs) are a subclass of timed automata with only one clock which resets at each transition.In this paper,we present an active learning algorithm for deterministic real-time automata(DRTAs) in both continuous-time semantics and discrete
Mobile edge computing(MEC)meets the high-bandwidth and ultra-low-latency requirements of mobile networks and improves communication reliability by reducing the networks\'load.In this study,we study a wireless powered MEC system,which includes an access
Although alpine ecosystems have been commonly recognized as sensitive to recent climate change,few studies have examined its impact on the long-term productivity of vegetation and soil erosion.Using paleoecological records,these two aspects were examined
Terahertz(THz)multiuser massive multiple-input multiple-output(MIMO)has been envisioned as a promising technique to support massive connectivity in next-generation wireless networks.However,due to ultra-high frequency band,the path attenuation of THz chan
Software defined networking(SDN)and network function virtualization(NFV)as new tech-nologies have shown great potential in improving the flexibility of resource management for network service provisioning.As traffic dynamics may cause violation of rigid s
In this paper,we investigate a user scheduling algorithm for massive multiple-input multiple-output(MIMO)systems over more general correlated Rician fading channels.To achieve low latency and high throughput,a new user scheduling algorithm based on deep l
Recent studies have revealed that long-term financial news can affect stock markets.However,previous research mainly focuses on modeling the short-term effects of financial news and suffers from the weak ability of quantifying the time-decaying influence