Current Virtual Machine (VM) live migration mechanisms only focus on providing a high availability service by offering minimal downtime to users. In this paper, we present a novel live migration technique called HSG-LM, which also aims to provide short waiting time to whoever is responsible for triggering the VM migration (e.g., the data center administrator). HSG-LM is implemented in the guest OS kernel in order to not rely on the hypervisor throughout the entire migration process. HSG-LM exploits a hybrid strategy that reaps the benefits of both pre-copy and post-copy mechanisms. Furthermore, HSG-LM integrates a speculation mechanism that improves the efficiency of handling post-copy page faults. From our evaluation on different real-world workloads (Sysbench, Apache, etc.), the results show that HSG-LM incurs minimal downtime as well as short total migration time. Moreover, compared with competitors, HSG-LM reduces the downtime by up to 55%, and reduces the total migration time by up to 27%.
No takes yet. Share an insight, caveat, or question.
Peng et al. (2013) studied this question.