HCI: Building a Lightweight IT Foundation for Disaster Recovery and Other Applications

Shenwan Hongyuan Securities Production Center has operated a virtualization resource pool based on "servers + active-active storage" since 2014, supporting over 90% of its production business systems. As usage time increased, the underlying hardware and software—including servers, networking, and virtualization—no longer met the company's evolving business demands. Issues included high failure rates, rising expansion and operational costs, and inadequate performance. Therefore, the user plans to introduce a hyper-converged architecture and has selected SmartX Hyper-Converged and DELL VxRail to host some business systems for a small-scale trial.

Users deployed asset management operations using hyper-converged solutions from SmartX software (on Inspur server hardware) and DELL (VxRail integrated hardware/software appliances). This approach established dedicated hardware clusters meeting high reliability, availability, and security requirements at minimal cost, earning recognition from business departments. The purely software-delivered SmartX hyperconverged system supports open compatibility across hardware and software, enabling heterogeneous configurations. It allows the use of different hardware within the same cluster—such as CPUs of varying generations and clock speeds—and facilitates flexible scaling by adding nodes as needed.

Users build domestic IT innovation clusters using domestically produced servers (Hygon, Kunpeng) and hyper-converged infrastructure, enabling centralized management alongside other clusters to achieve unified resource pool administration.

During the actual deployment process, users leveraged existing servers, adding only hard drives and a few other components to complete the disaster recovery cluster setup. The cluster network utilized 10 Gigabit Ethernet without purchasing new dedicated SAN storage switches. Final transaction system performance tests demonstrated that the disaster recovery system on this hyper-converged cluster achieved 58% of the primary production environment's performance.

Following the disaster recovery system's completion, it served not only as a static backup system but also progressively expanded the types of services hosted on the hyper-converged cluster. It simultaneously handled routine query tasks and pre-production testing for the live system, maximizing resource utilization.