Prove that nothing moved
In a major trading platform upgrade, such as a Murex upgrade, the functional scope stays the same but everything underneath changes. Proving equivalence is a reconciliation exercise, not a matter of opinion. EasyRec makes it repeatable, patch after patch.
An upgrade is proven at the end of the chain
It is not proven when trades can be booked. It is proven when the same outcomes come out at the end of the chain, for the full population, with every difference explained before sign-off.
- Trades & eventsBooking, lifecycle events
- ValuationMTM, P&L, cash flows, Greeks
- ProcessingWorkflows, messaging, payments, accounting
- Reporting tablesSuch as Murex Dynamic Tables: the reporting truth layer
- Reports & feedsDatamart, regulatory, finance, downstream
Upgrade regressions are data-dependent, which is why a handful of test cases cannot prove non-regression. EasyRec makes reconciliation the control layer across the complete end-to-end test.
Every patch restarts the validation
Business and IT teams are mobilised for months, environments are duplicated, and volume, data model complexity and platform-specific structures defeat Excel and generic comparison tools.
A small configuration change can alter thousands of reporting rows. The report still runs, while the figures underneath are wrong.
EasyRec does not reduce what has to be validated. It reduces the cost of validating it again and again.
Illustrative estimate. 200 is a floor: many upgrades involve far more reports.
600+ validations from one reusable configuration
Fifty reports or five hundred, the template is the same object. The first cycle builds the framework; every delivery after that reuses it.
- 01
Identify
Audit execution jobs to keep only relevant reports and their latest executions. The right scope from day one.
- 02
Classify
One taxonomy for any table: category (Accounting, FO, Risk), type (cash flows, P&L, positions, simulations, SQL) and priority.
- 03
Reconcile
One generic template loops over hundreds of reporting tables, with dynamic columns, conditional keys and conditional tolerances.
- 04
Automate
Batch validation scheduled every cycle, results in Excel and database, run-to-run comparison.
From first run to sign-off
Same inputs, one difference
Same reference database, same reference dates, same configuration, different platform version. Uncontrolled inputs are the most common cause of days lost on differences that were never regressions.
Schema first, then data
Deleted or renamed columns and reduced column sizes are detected before any data is compared. Schema non-regression is delivered in hours.
Tune on a sub-perimeter
Keys, conditional keys, noise columns and pattern tolerances are tuned on a sample. Each tuning cycle costs minutes, so the template converges in days.
Prioritise the investigation
The dashboard shows which reconciliations need attention. Start with fifty, spend the day on the handful that matter.
Re-run after each patch
No reconfiguration. Confirm the fix, and prove the patch changed nothing else: the hard part, fully automated.
Evidence for go / no-go
Run-to-run comparison, scope coverage per cycle, ageing of unexplained differences and exports for steering committees.
Run 1 → Patch → Run 2 → Patch → Run 3 → Sign-off
Automation turns validation into a repeatable capability
Less project effort
One configuration for the full reporting scope, flat as the estate grows.
Shorter cycles
Hundreds of controls re-run immediately after each delivery, not weeks later.
Experts on exceptions
Specialists investigate real regressions instead of extracting and eyeballing data.
Lower go-live risk
Full-population, repeatable and auditable evidence behind every decision.
Securing your next Murex or platform upgrade?
Show us your scope. We will show you how many validations one EasyRec configuration can cover.