No abstract

SUMMARYAs the virtualization technology becomes the core ingredient for recent promising IT infrastructures such as utility computing and cloud computing, accurate analysis of the internal behaviors of virtual machines becomes more and more important. In this paper, we first propose a novel I/O fairness analysis tool for virtualization systems. It supports the following three features: fine-grained, multimodal and multidimensional. Then, using the tool, we observe various I/O behaviors in our experimental XEN-based virtualization system. Our observations disclose that 1) I/O fairness among virtual machines is broken frequently even though each virtual machine requests the same amount of I/Os, 2) the unfairness is caused by an intricate combination of factors including I/O scheduling, CPU scheduling and interactions between the I/O control domain and virtual machines, and 3) some mechanisms, especially the CFQ (Completely Fair Queuing) I/O scheduler that supports fairness reasonable well in a nonvirtualization system, do not work well in a virtualization system due to the virtualization-unawareness. These observations drive us to design a new virtualization-aware I/O scheduler for enhancing I/O fairness. It gives scheduling opportunities to asynchronous I/Os in a controlled manner so that it can avoid the unfairness caused by the priority inversion between the low-priority asynchronous I/Os and high-priority synchronous I/Os. Real implementation based experimental results have shown that our proposal can enhance I/O fairness reducing the standard deviation of the finishing time among virtual machines from 4.5 to 1.2.

Recently, virtualized environments such as cloud computing and a virtual cluster are used popularly by lots of MapReduce applications to reap the benefits of low cost and flexibility. However, the I/O bottleneck of the virtualization software gives a burden especially for processing big data. To relieve the burden, we propose a novel burstiness-aware I/O scheduler. Our analysis has revealed that the I/O bottleneck is caused by I/O interferences among the bursty I/Os triggered by different virtual machines, especially when they execute the map and/or reduce tasks. The I/O interferences result in frequent context switches in the virtualization software and long seek distances in a disk. Our proposed I/O scheduler first detects I/O burstiness of a virtual machine on-line. Then, it schedules bursty virtual machines in a round-robin fashion so that a scheduled virtual machine utilizes most of I/O bandwidth without interferences. Real implementation based experiments haveshown that our scheduler can enhance the I/O performance up to 23% with an average of 20%.

No abstract

No abstract