Answer
Connectivity is often the deciding factor that gets skipped early in the evaluation. Cloud DR for IBM i depends on sustained bandwidth for initial data seeding and ongoing replication, and on network paths that can support 5250 and application traffic from wherever staff will actually be working during a declared event. A design that looks clean on a slide can fall apart when the team discovers that failing over means reconfiguring VPNs, DNS, and exit point security rules under pressure, with nobody having rehearsed that part.
Governance and cost tolerance matter just as much. Some organizations are contractually or regulator-driven toward keeping recovery infrastructure within specific jurisdictions or under direct control, which narrows or rules out managed cloud options. Others find that a cloud target actually costs less than maintaining a second physical LPAR and site, once hardware refresh and facilities costs are counted honestly. Buyers should ask cloud DR vendors for a real IBM i partition specification, not a generic cloud VM comparison, since IBM i licensing, storage architecture, and performance characteristics do not map directly onto commodity cloud infrastructure. A short, well-run cloud DR test is worth more than any vendor's capacity claims.