CSI baseline CSI-CARD-2026-10-22 · 22 October 2026. Limited to the registered source set. Application execution and proposed tests were not run; source findings do not establish runtime behaviour.
CardDemo: CSI overview
In this document
Assessment baseline: 22 October 2026. This assessment connects two inspected application paths to the business behaviour they support and the checks needed before changing them. It establishes four findings and five verification tasks. Every verification task is not run. No applicable CVE, deprecation notice or end-of-support date has been established. CARD-SRC-01 CARD-SRC-05 CARD-SRC-06 CARD-SRC-09
What the assessment establishes
The optional authorisation handler sends a correlated MQ reply before writing the corresponding IMS authorisation state. The base account-maintenance program rewrites the account before the customer and requests rollback if the customer rewrite fails. These are inspectable source relationships. Whether either path produces an inconsistent result in an installation still depends on the deployed executable, resource configuration and observed failure behaviour. CARD-SRC-02 CARD-SRC-03 CARD-SRC-04
Those relationships give a reviewer two specific places to start: the interval between an approval reply and its subsequent database write, and the recovery prerequisites of a paired account/customer update. The assessment does not establish a production incident, exploitability or an executed test result. CARD-SRC-02 CARD-SRC-03 CARD-SRC-09
The four findings
| Finding | Evidence status | Decision supported now |
|---|---|---|
| CARD-FIND-01: Authorisation reply before IMS persistence | Source order correlated with official MQ API semantics. | Review the reply/persistence boundary and adapt candidate C13-13 after binding the deployed baseline. |
| CARD-FIND-02: Paired update recovery prerequisites | Conditional data-integrity risk supported by source, sample configuration and CICS documentation. | Inspect effective file recoverability, then characterise a failure of the second write and concurrent modification. |
| CARD-FIND-03: Installed middleware versions missing | Documented gap in the inspected material. | Collect exact installed identities before assigning a CVE, deprecation notice or support deadline. |
| CARD-FIND-04: Separate evidence editions | Documented provenance gap. | Reconcile the retained source and public documentation before treating candidate scenarios as authoritative expected results. |
The first finding uses CARD-SRC-02, CARD-SRC-08, CARD-SRC-09, CARD-SRC-10 and CARD-SRC-11. The second uses CARD-SRC-03, CARD-SRC-04, CARD-SRC-09, CARD-SRC-12 and CARD-SRC-13. The inventory and provenance findings use CARD-SRC-01, CARD-SRC-05, CARD-SRC-06, CARD-SRC-07 and CARD-SRC-14.
Authorisation: compare the received reply with recorded state
COPAUA0C receives an MQ request, reads the relevant application information and makes the available-credit decision. It calls 7100-SEND-RESPONSE before 8000-WRITE-AUTH-TO-DB. The reply preserves the request correlation identifier. The later write updates authorisation summary and detail state in IMS. CARD-SRC-02 CARD-SRC-08
The source uses MQGMO-NO-SYNCPOINT for receipt and MQPMO-NO-SYNCPOINT for its MQPUT1 reply. IBM explains that a put outside syncpoint is immediately available and is not undone by backing out the unit of work. A requester could therefore observe approval while a later IMS write fails. That is a conditional consequence to investigate; actual reply receipt, database state and downstream effects have not been observed. The existing public candidate C13-13 addresses this precise boundary. CARD-SRC-02 CARD-SRC-09 CARD-SRC-10 CARD-SRC-11
Account maintenance: compare rollback intent with effective recovery
COACTUPC acquires account and customer records for update, compares original values, rewrites the account, and then rewrites the customer. A failed customer rewrite requests SYNCPOINT ROLLBACK. The sample definitions for ACCTDAT and CUSTDAT specify RECOVERY(NONE) and JOURNAL(NO). CARD-SRC-03 CARD-SRC-04
The sample settings do not determine the installed result. IBM describes rollback in terms of recoverable resources and documents the role of ICF catalog LOG attributes in determining recovery. The reviewer needs the effective installed definitions and catalog attributes, followed by observations of both records after a controlled second-write failure. C13-03 and C13-04 provide existing candidates for paired updates and concurrent-change handling. CARD-SRC-09 CARD-SRC-12 CARD-SRC-13
The five outstanding tasks
| Task | Required outcome | Status |
|---|---|---|
| CARD-TEST-01 | Characterise response receipt and IMS state when the later write fails, using C13-13. | Not run |
| CARD-TEST-02 | Inspect installed recoverability and characterise paired updates, using C13-03. | Not run |
| CARD-TEST-03 | Characterise stale-update rejection, using C13-04. | Not run |
| CARD-TEST-04 | Bind exact runtime, build and optional-module identities before external-notice matching. | Not run |
| CARD-TEST-05 | Establish the source baseline and authority of expected results. | Not run |
The public scenarios are proposals, not pass/fail evidence. The retained source archive and the public PDF edition remain separate baselines; their agreement on these paths does not establish complete equivalence. CARD-SRC-01 CARD-SRC-07 CARD-SRC-09 CARD-SRC-14
Scope of this report series
The assessment covers the optional MQ/IMS authorisation extension and the base account-maintenance path. It does not assume that the extension is deployed with every CardDemo installation. Technology names identify candidate dependencies, but the sample project's 1.0.0 badge does not identify an installed MQ, CICS, IMS, Db2, COBOL compiler or z/OS release. CARD-SRC-05 CARD-SRC-06
The component inventory, impact graphs, action plan and verification plan make the two traces reviewable. The evidence register records what was consulted and its access limits. The baseline report states what remains unknown. This is a dated assessment; it makes no claim of automated monitoring or continuous coverage.
This Wayfinder CSI assessment is limited to the registered source set. Findings, inference, recommendations and evidence gaps remain distinct. It does not establish a completed migration, production suitability or functional equivalence.
Independent Harten analysis of AWS Samples CardDemo, a public demonstration application. No AWS endorsement is implied. Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved. Apache 2.0 source licence. The source-licence reference does not establish the revision analysed by the retained reports.
Inspect the evidence and external lookups · Read the existing application documentation edition.