To guarantee the availability of data centers, ZTE provides multilayer all-round reliability solution for single data center to crossing multiple data centers, to automatically switch services without any customer perception. It includes the following two parts:
1.Backup Recovery Scheme
Based on the mainstream open source distributed storage project, Ceph, ZTE backup recovery system provides backup capability from cloud platform layer to VM layer, including multiple kinds of networking architectures and rich backup functions, to solve the complexity caused by diversified backup demands. In addition, in combination with VM image snapshot function, it implements rapid backup and recovery of application layer, and provides total, incremental and other backup modes for different application requirements.
2.Disaster Recovery (DR) Schemes
With deep understanding of DR construction for multiple data centers and rich technology practice experience, ZTE provides the following DR schemes of different levels as well as a visual management system to improve the management capability of DR services.
1)Geographic active/standby DR: A production center and a DR center are respectively built at two places, and each center is deployed with one set of TECS cloud operating system (OS). Data of the production center is synchronized to the DR center through the synchronization function of storage disk
array or Ceph cluster. Thus, when the production center fails, the DR center is started rapidly, to recover services in the shortest time.
2)Alive-alive DR: A production center and a DR center are built in the same city, and one set of TECS cloud OS is used to manage their resources. The two centers are connected by L2 and support cross-site VM hot migration and rebirth, so as to automatically switch services without any customer perception, RPO=0, RTO≈0.
3)2-site 3-center DR: Combining the previous two schemes, it provides multilevel protection for data and services of the production center, guaranteeing service zero-interruption and data zero-loss.
Disk image technology based on shared volume is adopted to guarantee zero-loss of user data, RPO=0.
2. Cross-DC VM rebirth technology is adopted to guarantee rapid service recovery, RPO≈0.
3. Scalability-based backup recovery components are adopted to provide customers with customizable backup recovery content.
4. Multiple kinds of back modes, such as full backup, incremental backup, and differential backup, are adopted to meet different customer demands.