Estate map
Artifacts, online and batch classification, copybooks, jobs, transactions, metrics, and exceptions.
Starlight Discovery
See the estate before you change it.
Starlight Discovery is MSS International's evidence-first assessment platform for IBM and Unisys modernization programs. It turns source, jobs, data definitions, dependencies, and business-rule evidence into a governed workbench for decisions.
What it does
Discovery that produces usable engineering evidence, not just inventory.
Starlight is built for customer-site execution and high-consequence source estates. It preserves source traceability while producing the structured evidence needed for planning, validation, modernization, and assisted transformation.
Relationship graph
Calls, orchestration, data access, IPC, services, PROC usage, and bounded impact neighborhoods.
Business evidence
Data dictionary, rule candidates, source-linked statements, control flow, data flow, slices, and pseudo-code.
ONEM
Starlight Discovery
Snapshot complete
SYS/SBFreal artifact
SYS/SBF/VARScopy evidence
BLLIB/MOVELISTS37,668 rules
CUSTOMER-MASTERread/write impact
Current Object
Every view stays grounded in source path, line number, output file, confidence, and warnings.
Discovery Workbench
Navigate from source to evidence and back again.
Starlight treats each artifact, rule, data item, graph node, and source line as an object with context. Teams can inspect the neighborhood around a program, trace a data item, review business-rule candidates, and assemble bounded evidence for deeper analysis.
Modernization flow
From customer source to governed modernization insight.
Run deterministic discovery across IBM z/OS, Unisys MCP, and Unisys OS 2200 estates.
Import JSON evidence into a queryable PostgreSQL knowledge store with immutable snapshots.
Use the workbench to inspect source-linked evidence, exceptions, dictionary changes, and graph context.
Feed reviewed evidence into planning, pseudo-code, AI-assisted interpretation, and migration workflows.
Why it matters
AI can accelerate modernization. Starlight keeps it accountable.
Provenance first
Every meaningful fact links back to generated evidence, source files, logical names, and line references.
Confidence is explicit
Observed, derived, inferred, and curated findings are separated so teams know what is known and what needs review.
Built for secure delivery
Offline runs, customer-site execution, snapshot history, audit, and source-provider boundaries are part of the model.
Starlight Discovery