Post-trade architecture

Architecture is the difference between a post-trade system and a collection of reconciliations.

Devancore is built around controlled lifecycle records, unified position state, rail-aware settlement evidence, and embedded controls across traditional and digital post-trade workflows.

Post-trade infrastructure has to do more than store trades. It has to explain what happened, which record is authoritative, how the position changed, what settlement state applies, who approved the action, which control operated, and what evidence can be produced later.

Legacy operating models often split those answers across product systems, asset-class ledgers, settlement tools, reconciliation files, reporting databases, and supervisory workflows. Each new product or rail adds another place where the firm must reconcile records before it can trust the result.

Devancore's architecture starts from a different assumption: the lifecycle record should carry the operating truth. Trade capture, enrichment, settlement instruction, finality status, position movement, control approval, exception handling, and reporting evidence should be connected by design, not reconstructed after the fact.

This page explains the architecture model at the operating level. It does not disclose internal schemas, proprietary algorithms, infrastructure topology, or implementation details. The point is to show how the public product architecture supports post-trade records, controls, settlement state, reconciliation, and reporting evidence without turning every asset class or settlement rail into a separate operating stack.

Devancore provides software infrastructure for post-trade records, workflows, controls, reconciliation, and reporting evidence. It does not execute trades, custody assets, clear transactions, supervise firms, provide legal advice, or guarantee regulatory compliance.

Swimlane diagram showing trade, settlement, and finance flows from input through lifecycle event, position state, control, and evidence columns, converging on Devancore normalize, record, book, control, and preserve stages.

Lifecycle records

The lifecycle record is where workflow, audit, and reporting begin.

An event-centered operating model records the sequence of post-trade work as it happens, so later views of position, settlement, control, and reporting evidence stay tied to the underlying lifecycle.

  • Post-trade work is a sequence of state changes

    A trade is captured. Account and instrument context is applied. Allocation and confirmation steps occur. A settlement instruction is created. Status changes arrive. Exceptions are routed. Controls operate. Positions and cash move. Records are preserved.

  • Disconnected systems create reconciliation debt

    If those steps live in disconnected systems, the firm eventually has to reconstruct the story. That reconstruction becomes reconciliation debt: matching logs, positions, messages, files, approvals, and reports after the operating work has already happened.

  • One controlled history, multiple views

    An event-centered model keeps the operating sequence visible. It treats lifecycle activity as the foundation for the downstream views that finance, operations, compliance, supervision, and audit need. The operating consequence is practical. A firm should not have to choose between a workflow log, an audit log, and a reporting base. Those views can be produced from the same controlled lifecycle history when the architecture keeps the record connected from the start.

  • Regulatory evidence depends on the record

    This is also why architecture matters for regulatory evidence. Books and records, audit trail, supervisory review, and examination response all depend on the ability to reconstruct what happened without guessing which system was authoritative at which moment.

What the lifecycle record needs to preserve

Record element Why it matters
Source and timestamp Establishes when the event entered the operating model
Account and instrument context Explains which book, position, and obligations are affected
Actor and workflow state Shows who acted and what step the workflow reached
Control and approval state Preserves maker-checker, role, and escalation evidence
Settlement and finality state Links the trade lifecycle to rail-specific settlement evidence
Position and cash impact Connects operational work to current state and reporting
Exception outcome Preserves break resolution, repair, and supervisory review

Ordered evidence sequence from source event through enriched lifecycle record, control state, position impact, settlement status, exception outcome, to reportable evidence.

Position state

Asset class and rail should be attributes of the record, not reasons to create another stack.

A unified operating record lets a firm view positions, cash, obligations, settlement state, and evidence across instruments without building a separate post-trade system for every product or rail.

  • Every sub-ledger creates a reconciliation boundary

    The problem is not only technical complexity. It is operating ambiguity: which system owns the position, which one owns the settlement status, which one owns the cash impact, and which one produces the evidence?

  • Dimensions, not duplicate stacks

    Devancore's architecture treats instrument type, asset class, account, market, custodian, and settlement rail as dimensions of the operating record. That allows the same post-trade model to support traditional securities workflows, digital asset workflows, tokenized instrument workflows, and hybrid settlement workflows without implying that every instrument has identical legal or operational treatment. The point is not to flatten regulatory differences. The point is to avoid duplicating the operating record every time a new instrument or rail appears.

  • Preserve distinctions without duplicating books

    This model is especially important in hybrid environments. A tokenized instrument, traditional security, stablecoin payment leg, DTC position, custodian balance, or ISO 20022 instruction may produce different evidence and legal consequences. The architecture should preserve those distinctions without forcing the firm to operate separate books that need to be reconciled back together.

What a unified position model should support

Capability Operating purpose
Instrument-agnostic record design Avoids separate operating stacks for each asset class
Account-aware position state Keeps customer, proprietary, fund, and omnibus context attached
Cash and securities linkage Connects position movement to payment, reserve, and reporting context
Settlement state Shows unsettled, instructed, matched, failed, repaired, and final states
Control state Preserves approval, exception, and supervisory evidence
Reporting view Lets finance and operations read from the controlled operating history

Before and after comparison: a fragmented post-trade stack with separate asset-class records and assembled evidence versus a unified operating record where rail and asset class are attributes, controls run inside workflow, and reporting reads from controlled history.

Rail-aware finality

Settlement finality should be modeled explicitly, not inferred from a generic settled flag.

Different rails produce different forms of confirmation. The architecture has to preserve the source, status, timestamp, and evidence that support finality.

  • Finality is not one universal signal

    A DTC or NSCC workflow, custodian confirmation, bank cash movement, SWIFT or ISO 20022 message, and blockchain event can each provide different evidence about the state of an obligation.

  • Context matters

    A useful post-trade system records the rail, instruction, source, actor, timestamp, finality state, exception state, and evidence reference — rather than collapsing those signals into a generic status without context.

  • Where operations and regulation meet

    This matters because finality is where operations, finance, customer protection, and reporting meet. A settlement state can affect position availability, cash movement, failed trade aging, reserve inputs, capital support, reconciliation, supervisory review, and client reporting.

  • One disciplined record for every rail

    The architecture principle is simple: the rail can vary, but the operating system still needs one disciplined way to record settlement state and evidence.

Finality evidence by source

Source What it may provide What the operating record should keep
CSD / clearing infrastructure Settlement status, netting output, participant confirmation Settlement event, status, account, position impact
Custodian or bank Statement, advice, cash or custody confirmation External evidence, reconciliation input, cash or custody state
ISO 20022 / SWIFT Instruction, status, statement, cash movement messages Message reference, instruction state, settlement or cash event
Blockchain or wallet Transaction hash, block, contract event, balance evidence Digital rail event, confirmation source, finality state
Exception workflow Break, repair, approval, escalation Exception case, owner, control state, closure evidence

Table mapping DTC and NSCC, SWIFT and ISO 20022, custodian and bank, and blockchain sources to their messages, operational purpose, and the operating records they produce.

Different rails produce different finality evidence

Devancore · message matrix

Rail Message Purpose Record
DTC / NSCC Settlement status, CNS output, participant confirmation Securities movement and fail state Settlement event and position update
SWIFT / ISO 20022 sese, semt, camt, MT equivalents Instruction, custody, and cash evidence Instruction, custody event, cash movement
Custodian / bank Statement, advice, balance, confirmation External account evidence Cash or custody reconciliation input
Blockchain / wallet Transaction hash, block, contract event, balance Digital rail evidence Digital settlement event and finality state

Embedded controls

Controls should operate inside the workflow, not after the workflow fails.

Maker-checker, role-based access, exception review, and audit evidence are architecture concerns because they determine whether the system can prove how work was controlled.

  • Overlays weaken evidence

    Post-trade controls are often treated as overlays: a separate approval log, a manual review queue, an audit export, or a compliance report generated after the business process is complete. That approach weakens the evidence. If the control does not operate inside the workflow, the firm has to prove after the fact that the work was reviewed, authorized, escalated, and resolved correctly.

  • Controls travel with the event

    Devancore's public architecture model treats controls as part of the lifecycle record. Approval, role, exception, reviewer, timestamp, and control outcome should travel with the operating event they govern.

  • Architecture and regulation share one question

    This is the bridge between architecture and regulation. Rules about books and records, customer protection, financial responsibility, and supervision all rely on the same practical question: can the firm show what happened, who controlled it, and which record supports the conclusion?

  • Workflow evidence stays with the record

    Workflow evidence is not a separate reporting task when controls are embedded in the architecture from the start.

Control evidence the system should create

Control evidence Operating purpose
Maker-checker approval Shows that sensitive actions received second-person review
Role-based access state Shows that only authorized users could perform the action
Exception ownership Shows who owned repair, escalation, and closure
Review history Shows supervisory or operational review sequence
Immutable audit trail Preserves the record used for examination or audit response
Reporting support Connects control evidence to finance, compliance, and management outputs

Integration model

Modern post-trade architecture has to coexist with the market infrastructure firms already use.

The architecture should connect to execution, settlement, custody, bank, reporting, and digital rail sources without forcing a rip-and-replace operating model.

  • Transformation rarely starts from a blank slate

    Post-trade transformation rarely starts from a blank slate. Firms already have OMS, EMS, custodians, banks, clearing relationships, reporting systems, risk systems, data warehouses, and manual procedures. A useful architecture has to connect into that environment.

  • Controlled operating record as the integration anchor

    The goal is not to replace every system at once. The goal is to create a controlled operating record that can ingest activity, normalize lifecycle state, preserve evidence, and expose reliable downstream views.

  • Clear authority for external activity

    Architecture should reduce the number of places where records diverge. That does not require pretending external systems disappear. It requires a clear authority model for how external activity becomes lifecycle evidence inside the operating book. Connectivity is how external market activity becomes controlled lifecycle evidence inside the operating record.

Connectivity categories

Connection type Why it matters
FIX Execution, allocation, and trading workflow integration
ISO 20022 / SWIFT Settlement instruction, status, custody, cash, and statement messaging
Custodian and bank data External account, balance, position, and confirmation evidence
APIs Controlled downstream access for finance, risk, reporting, and operations
Digital rail events Wallet, transaction, block, contract, and token movement evidence where applicable
Reference data Account, instrument, SSI, counterparty, and classification support