| Application availability research is an important means to ensure application software stable operation, but virtualization-based high availability solutions have two main drawbacks:1) the current application HA-aware solutions need to modify lots of application code which leading to the software complexity is relatively high; 2) the VMM-based (virtual machine monitor) availability solutions have serious performance loss. Therefore, we proposed the container-based availability solution, and verified the feasibility which took data integrity protection software for case study.The main work is as follows:1. Research the existed applications availability and summarize the advantages and disadvantages of these studies; research two key core technologies for realization of Linux container virtualization:Cgroup and Namespace.2. An availability solution of separated data and application is proposed based on Linux container. Application is insulated in the container, and the application data stored in the data volume where all containers are visible. High availability agent ensures the integrity of the application data, and the container manager monitor the application and the container respectively. When there is abnormal operation, HA-agent recover the data and restore the normal operation of the application and the container.3. According to the data integrity protection requirements of Linux security subsystem, we design and realize Linux data integrity protection software. By extracting the data file attribute set and regularly monitored attribute set snapshot, we can detect inconsistencies and then process the damage.4. Based on the proposed availability solution, we enhanced the software availability for the above-mentioned data integrity protection. Linux containers lightweight virtualization helps reduce performance overhead, and the solution do less to modify the software so as to reducing software complexity. |
PDF Full Text Request