A participant should be able to reach another payment rail through a controlled extension of its banking integration, with each network’s responsibilities kept explicit.
Every new rail reaches the bank’s integration boundary
A bank can connect to a payment network and still face a difficult integration problem. Its Core Banking system must understand incoming requests, verify accounts, make posting decisions, follow outcomes and handle returns. Those responsibilities reach far beyond sending a message to a switch.
Adding a second rail can multiply that work. Different message profiles, signatures, destinations and response sequences can spread into the bank’s integration layer. Over time, each new network risks becoming another implementation of the same banking capabilities.
SPS Connect, previously known as SIPS Connect, provides a common participant-facing foundation for domestic SIPS and PAPSS. The bank exchanges JSON with SPS Connect. The connector validates, maps and routes those requests into signed ISO 20022 messages for the selected network service. Incoming traffic follows the reverse path into the bank’s configured APIs.
The architectural value is a stable place to connect banking responsibilities to network requirements. A shared interface reduces repeated integration work while retaining the context needed to process each rail correctly.
Integrate the participant around banking capabilities
The participant contract centers on account verification, payments, status and returns. For outbound operations, the bank calls the Gateway API. For inbound operations, SPS Connect calls the bank’s configured verification, transfer, completion-notification and return endpoints.
This is a bidirectional integration. The participant initiates instructions and also receives work that requires its own decision. An incoming payment becomes a transfer request to the bank; an incoming beneficiary enquiry becomes an account-verification request.
The bank keeps control over account checks and ledger processing. SPS Connect translates the request, communicates the bank’s decision and delivers subsequent network outcomes. Endpoint URLs are configurable, and adapter mappings accommodate the bank’s JSON contract. Integration still requires agreed fields and behavior, but the bank does not have to organize its internal APIs around the network’s XML structure.
One participant boundary, distinct rail processing
Instructions and outcomes travel in both directions. PAPSS connectivity continues through its dedicated service or gateway; the participant retains account and ledger decisions.
- Participant bankJSON requests, account verification and transfer decisions
- SPS ConnectValidate, map, route, sign and track
- Selected rail serviceDomestic SIPS processing or the dedicated PAPSS service
- Network outcomeCorrelated results returned through notifications and lookups
Managing different bank integrations through one connector
A common participant contract does not require every bank to expose identical internal APIs. Banks operate different Core Banking platforms and middleware. SPS Connect accommodates those differences through deployment-specific endpoint configuration and request/response mappings for the bank being integrated. This describes a configurable connector boundary; it does not imply a central self-service platform that manages every bank from one console.
The integration configuration identifies where verification, transfer, completion and return requests should go. JSON mappings associate the bank’s field paths with the connector’s internal contract. The current integration guide permits customization of those external paths while requiring the internal fields, data types, null and empty-value behavior, and endpoint mapping names to remain intact. Flexibility has a defined contract.
This separates a bank’s internal API shape from the network interface. If the bank changes a field path or moves an endpoint, the corresponding configuration or mapping may absorb that change without altering the network message contract. A change in business meaning still requires agreement and validation; a mapping cannot make incompatible payment semantics equivalent.
Timeout settings also form part of the integration boundary. A bank API that does not respond in time creates a condition that must be recorded and resolved under the relevant workflow. It does not authorize the connector to invent a bank decision.
Maintaining this boundary requires disciplined change control. Endpoint and mapping changes should be versioned, tested against representative requests and responses, and supported by rollback procedures. Credentials, certificates and their renewal need assigned owners under the agreed authentication profile. These are operating requirements for a maintainable deployment, rather than a claim that SPS Connect provides an automated management console for every task.
Enabling another rail should follow the same discipline: validate the additional fields, capabilities and result handling, then run regression checks on the existing domestic path. The established banking capabilities can be reused, but successful reuse must be demonstrated through integration tests.
The same payment interface preserves the selected rail
A participant can submit either rail through POST /api/v1/Gateway/Payment. The instruction explicitly selects SIPS or PAPSS alongside the payment’s amounts, currencies, accounts, parties and destination details.
SPS Connect authenticates the caller, validates the instruction and selected rail, and applies the appropriate processing path. The rail determines the configured destination, message mapping and signing profile. A common API therefore provides a common entry point while preserving separate network handling.
For PAPSS payments, the connector also checks signed discovery and readiness information to validate destination capabilities. PAPSS must be enabled. If the requested PAPSS path is unavailable, the instruction does not silently become a domestic payment.
That behavior matters to the participant’s business intent. Choosing a rail is part of choosing how an instruction will be processed. The connector must preserve that choice throughout validation and dispatch.
Operating principleStandardize how the participant connects; preserve how each network processes the instruction.
One incoming endpoint still requires distinct trust decisions
Domestic SIPS and the PAPSS service deliver signed ISO XML to the shared POST /api/v1/Incoming endpoint. Sharing an address simplifies the entry surface, but every message must still establish its authority and processing context.
Before dispatch, the security boundary must establish message integrity, certificate and sender authorization, recipient identity, and compliance with the applicable network security profile. A valid XML signature alone does not establish that its signer is authorized to represent the claimed participant or network. The verified and authorized network context selects the applicable handlers. The ISO message type then identifies the operation: acmt.023 and acmt.024 for beneficiary verification requests and results, pacs.008 for payments, pacs.002 for status reports, and pacs.004 for returns.
These decisions answer different questions. Network identity establishes the context in which a message is trusted. Message type establishes what operation it represents. Currency or country alone cannot safely establish the rail.
This distinction lets the connector share infrastructure without treating every incoming message as an interchangeable instruction.
Follow an incoming PAPSS payment into the bank
Incoming processing depends on the rail and configured processing mode. The participant performs the required account or posting operations and returns the appropriate acknowledgement or business decision. Domestic authorization, credit-transfer advice and Store-and-Forward paths should not be read as one universal synchronous acceptance workflow.
The PAPSS incoming-payment path described here makes the responsibility boundary concrete. SPS Connect validates the message, stores it and checks for duplicates. It maps the instruction into the bank’s JSON contract and calls the configured transfer endpoint.
The bank checks the account and applies its own processing rules. It returns acceptance or rejection with a reason. SPS Connect persists the decision and sends the corresponding signed response through the PAPSS service path. A later network outcome is delivered to the bank as a separate mapped notification.
The transfer call and the notification serve different purposes. The transfer endpoint obtains the bank’s business decision. A notification callback communicates a subsequent event; it does not replace that decision-making call.
A timeout or an unrecognized bank response cannot be promoted into acceptance. Initial bank acceptance, customer-account crediting, network completion and interbank settlement are distinct facts. A successful beneficiary credit does not by itself prove that the related interbank settlement obligation has been finally discharged. Preserving those distinctions gives the bank an accurate basis for customer communication and operational follow-up.
The integration continues after submission
For an outbound PAPSS instruction, an initial response can report RECEIVED_AND_DURABLY_ADMITTED. That means the request has been admitted for processing. It does not establish settlement.
Later results arrive through authenticated callbacks. The participant can also retrieve progress through operation and payment lookups using the request-message or transaction reference. Those references connect the original instruction to its subsequent events.
Callbacks can be delivered more than once, so the bank must process them idempotently. A repeated notification should update or confirm the same business record without causing another posting. Successful notification delivery and successful payment processing also need to remain distinguishable during recovery.
A multi-rail participant contract must therefore describe more than request fields. It must explain acknowledgement, acceptance, completion, uncertainty and duplicate delivery well enough for the bank to act correctly at each stage.
Effectiveness needs evidence beyond the architecture
A clear integration boundary explains what SPS Connect is designed to do. Establishing how effectively it does that requires evidence from onboarding, testing and operations. Configurable mappings alone do not establish shorter implementation times, easier support or reliable production performance.
Useful evidence includes the time and engineering effort needed to onboard a bank, contract-test results for that bank’s mappings, duplicate and timeout scenarios, certificate-renewal exercises, and domestic regression results when PAPSS is enabled. Incident records can show whether correlation references let support teams locate a failed exchange and recover it without repeating the financial action.
The companion PAPSS delivery account records end-to-end testing, UAT certification and PAPSS SIT certification for that integration. Those are scoped qualification milestones. They do not independently demonstrate every bank configuration, comparative cost savings or sustained production reliability. This article makes no quantitative claim about those outcomes.
A shared foundation for participation
Integrating once is best understood as establishing a reusable banking integration foundation. Another rail can still require onboarding, additional fields, capability checks and changes to participant rules. The benefit is that those additions enter through an established contract rather than a completely separate network implementation inside the bank.
SPS Connect concentrates rail-specific routing, message protection, translation and workflow handling at the connector boundary. The dedicated PAPSS service retains its external network responsibilities. Participants retain their account and ledger decisions, alongside the funding, liquidity and settlement obligations associated with their participation or sponsor arrangements.
This is the design principle I want to emphasize: broader network reach should build on a participant’s existing capabilities. A bank should be able to understand what it must decide, what the connector handles and what remains a network responsibility. SPS Connect makes that division the foundation of multi-rail integration.
Connect the participant through common banking capabilities, retain explicit rail context, and follow every instruction through its decisions and outcomes.
Inspect the basis for this note.
These links point to public implementation evidence or authoritative documentation used for the claims above. The interpretation is mine; institutional code and ownership remain with their respective owners.
- 01Public participant integration foundation
The version-pinned SIPS Connect architecture documents the Core Banking adaptation and ISO 20022 orchestration foundation. The historical project name remains in the repository URL.
Open source evidence - 02The PAPSS delivery account
My companion article explains the dedicated PAPSS gateway, separate trust boundaries, participant responsibilities and qualification scope. It provides the wider architecture behind this participant-facing account.
Open source evidence - 03Duplicate handling and recovery background
The public financial safety documentation establishes the existing SIPS Connect foundation for identity reservation, duplicate control and recovery. It is background rather than independent verification of the newer multi-rail implementation.
Open source evidence