Research Of Optimizing Virtual Machine Checkpoint

As an important feature of virtualization, checkpointing can save the whole executionenvironments status of the running virtual machine (VM) for later restoration and ensurethat the VM continues to run after completion. It provides great benefits for systemmaintenance, fault-tolerant, testing and debugging in modern clusters and data centers.However, the system-level checkpointing needs to save a large number of data to the disk.Moreover, the overhead grows linearly with the increasing size of virtual machine memory,which leads to disk I/O consumption disaster along with poor system scalability.To target this, a novel fast VMs checkpointing approach is proposed. By studying therun-time memory characteristics of different workloads, the approach counts the dirtypages in a fine-granularity manner (the number of8bytes), instead of the conventionalmethod (the number of pages); the approach utilizes dirty page logging mechanism torecord the dirty pages and dirty bitmap to mark them; accordingly, a delta memorycompression mechanism is implemented to eliminate redundant memory data incheckpointing files. To locate the dirty data in dirty pages, the approach utilizes twomechanisms: by analyzing the distribution characteristics of dirty pages in dirty bitmap, afast dirty bitmap scanning method is proposed to locate the dirty pages, and take amulti-threading data comparison mechanism to locate the real dirty data in one page.The experimental results show that compared with Xen’s default system-levelcheckpointing algorithm, the approach can reduce70.54%of checkpointing time onaverage with1GB memory size and achieve better improvement for VMs with largermemory configurations; the approach can reduce52.88%of the checkpointing data size onaverage compared with Remus’s incremental solution which is in page granularity.Compared with default dirty bitmap scanning method in Xen, the scanning time of theapproach is decreased by91.13%on average.