The lead time till failure of modern mainframes is estimated at 12-15 years. Such reliability is the result of almost 60-years of improvements. The VM / ESA operating system development team spent 20 years troubleshooting and as a
Increased system stability. Mainframes can isolate and fix most hardware and software errors by using the following principles:
Data Integrity. Mainframes use memory with error correction. Errors do not lead to the destruction of data in memory or data waiting to be output to external devices. Disk subsystems, built on the basis of RAID-arrays with hot swapping and built-in backups, protect against data loss.
Workload. The mainframe workload can be 80-95% of their peak performance. The mainframe operating system will process everything at once, and all applications will work closely together and use common software components.
Throughput. Mainframe I/O subsystems are designed to work in an environment with the highest workload of I/O data.
Access to data. Because the data is stored on a single server, application programs do not need to collect source information from multiple sources, no additional disk space is required for their temporary storage, there is no doubt about their relevance. It requires a small number of physical servers and much simpler software. All this, in total, leads to an increase in the processing speed and efficiency.
Protection. The built-in security features, such as cryptographic devices and Logical Partition, and operating system protection, supplemented with RACF or VM: SECURE software, provide reliable protection.
User interface. The user interface of mainframes has always been its weakest point. Now it has become possible for mainframe applications to provide a modern web interface in the shortest possible time and at minimal cost.