Rocket U2: priced on users and devices, not on MSU.

Rocket UniVerse and Rocket UniData are the two databases that make up the Rocket U2 family, a MultiValue (post-relational) data platform that has carried mission critical applications for decades across banking, retail, distribution, and government. UniVerse traces to the PICK lineage and UniData to a parallel MultiValue heritage; Rocket Software (formerly the steward of the Micro Focus and ASG portfolios it has consolidated) now owns and maintains both, along with the U2 development tools and the free MultiValue Integration Server. They run on UNIX, Linux, and Windows rather than as z/OS sub-capacity products, but they sit squarely inside the legacy enterprise estates we are routinely asked to baseline alongside the mainframe.

U2 is licensed primarily on concurrent access rather than processor capacity. The common metric is the number of simultaneous users or devices authorized against the data server. Device licensing for the Server Edition lets a buyer tune the number of connections allowed per device, such as a single user's PC, and connection pooling for stateless and web workloads is authorized separately so it does not consume data server licenses. The U2 clients ship at no additional cost. Because the metric is seats and connections rather than MSU, a U2 entitlement is not measured through SCRT and is not a mainframe MLC or IPLA line; the cost moves with the user and device count and the priced editions and connectors enabled.

Directional and pattern level. Editions, device rules, and connector terms evolve, so confirm the current model in your own U2 schedules before modeling a renewal.

The first cost driver is the licensed user or device count, which tends to be set high and rarely revisited, so it often sits above true concurrent demand once an application matures. The second is the set of priced editions, options, and connectors enabled on top of the base database, which accumulate over years of project work and are seldom retired when the project ends. The third is environment sprawl: test, training, and disaster recovery copies that the agreement may scope more narrowly than production. Maintenance is charged as a percentage of license value, so any padding in the base entitlement compounds every year it is renewed.

U2 exposure sits in the gap between what is deployed and what is contracted. Common traps we see at pattern level:

A U2 estate is long lived, which means years of accumulated entitlement to reconcile. The five levers that pay:

MultiValue is a niche, and migrating a UniVerse or UniData application to a relational database or a different MultiValue platform is a real project with real risk, because the data model and the application logic are tightly coupled to U2 behavior. That makes wholesale replacement a slow lever, not a fast one. The more reliable position is scope discipline and a credibly prepared modernization path held in reserve, since Rocket itself sells the modernization tooling and has an interest in keeping the estate inside its portfolio. Price any migration honestly, including retraining and parallel running, before treating it as leverage.