What Is a Mainframe?
A mainframe is a high-reliability, high-throughput computer system designed for mission-critical transaction processing. In 2026, the term almost always means IBM z Systems hardware running z/OS — the descendants of the System/360 line IBM introduced in 1964. Mainframes process the overwhelming majority of the world's financial transactions, insurance claims, and government records.
- First shipped
- 1964 (S/360)
- Current vendor
- IBM z Systems
- Availability target
- 99.999%
- Global footprint
- ~10,000 active systems
The hardware: from System/360 to z16
IBM introduced the System/360 in 1964 — a family of compatible machines covering a wide range of price points, all running the same software. That compatibility promise is still alive: code compiled for a 1970s System/370 can still run on a 2024 z16, because every new generation has preserved backward compatibility at the instruction-set level.
Modern z Systems hardware is specialized. Redundant everything — dual power, redundant cooling, hot-swappable processor books, error-correcting memory that can replace a DIMM without a reboot. Mainframes are designed to keep running when parts fail.
The operating system: z/OS
z/OS is IBM's flagship 64-bit mainframe operating system, descended from OS/360 through MVS, OS/390, and finally z/OS. Its core subsystems — CICS for online transaction processing, Db2 for relational data, IMS for hierarchical and transactional data, RACF for security, JES for batch — are the real platform. Most mainframe applications are actually COBOL or PL/I programs running inside CICS or as JES batch jobs.
Why mainframes persist
- Reliability. Five-nines availability isn't aspirational — it's contractual. Major institutions measure mainframe downtime in minutes per year.
- Throughput. A single z16 can handle tens of thousands of transactions per second while running hundreds of concurrent workloads under z/OS Workload Manager.
- Sunk cost of correctness. Regulators have signed off on the applications that run on these systems for decades. Replacing that sign-off is expensive.
The modernization pressure
At the same time, running on the mainframe is expensive. IBM software licenses scale with MIPS consumption, hardware refreshes cost tens of millions, and the pool of engineers who can maintain z/OS applications is shrinking. Modernization decisions usually come down to which workloads to move and how — not whether to move anything at all.
Hypercubic's approach: capture the knowledge before it retires (HyperTwin), document the code so any engineer can pick it up (HyperDocs), run operations with AI help (Hopper), and modernize what should be modernized with formal correctness proofs (HyperLoop).
Related terms
Explore COBOL, JCL, CICS, Db2, IMS, RACF, VSAM, LPAR, and the full mainframe glossary.