...
The core detection capability within the platform is distributed across 3 three distinct steps in the end-to-end evaluation flow.
...
Changes to the rule and typology scope of the evaluation (a - network map)
In the rule processors:
Changes to the parameters that influence the rule processors' behavior (b - rule config)
Changes to the result bands that classify the rule processors output' outcomes (b - rule config)
In the typology processor:
Changes to the rule result weightings (c - typology config)
Changes to the typology threshold (c - typology config)
In this document, we will discuss how the various configuration documents are expected to be updated to influence evaluation behavior.
...
The platform processes configurations in a specific order to evaluate an incoming transaction. Starting with the Channel Router & Setup Processor (CRSP) that interprets the network map for routing, then following with the rule processors that interpret their individual rule configurations to determine how to evaluate the transaction, and then concluding with the typology processor that uses a variety of typology configurations to summarize rule results into typologies (fraud or money laundering scenarios).
...
The development cycle for the platform processors and their associated configurations follow a slightly different flow. The development and configuration process follows somewhat loosely cascading dependencies amongst the configuration documents: typologies rely on rules, and the network map that defines routing relies on typology-and-rule structures.
...
Rule results roll up into typologies through a typology configuration. One would typically only start composing a typology once the target rules have been developed and deployed, though sometimes rules may be added to, or removed from a typology, and through that its configuration, an existing typology. Without a view of the target rules and their configurations, blindly composing a typology would be very difficult, so this step usually follows the completion of the rules.
...
id
identifies the specific rule processor and its version related to that will use the configuration. It is recommended that the rule processor “name” is drawn from the source-code repository where the rule processor code resides, and the version should match the semantical version of the source code as defined in the source code repository.cfg
is the unique version of the rule configuration. Multiple different versions of a rule configuration can co-exist simultaneously in the platformdesc
offers a readable description of the rule
...
Parameter | Description |
---|---|
| The time-frame that defines the intervals into which a histogram is partitioned. Some rules perform a statistical analysis of behaviour over time and partitions the historical data into a histogram. This parameter defines, in milliseconds, the time-frame of each interval. |
| The maximum number of records to return in the query. This parameter limits the number of results that can be returned from the database. |
| A time (in milliseconds) that limits the maximum extent of a historical query. A query with a value of 86400000 would only look up messages received within the last 24 hours. |
| The least number of transactions required for the rule processor to produce a result. Some statistical algorithms required at least a certain number of data-points to be able to render a useful result. If the minimum number of transactions cannot be retrieved, the rule processor will raised a non-deterministic exit condition. |
| A margin of error for an evaluation against a threshold. With a tolerance of 0 (zero) the match against a target value would have to be exact, but with a tolerance value of 0.1, the match could be in a range either 10% below or above the threshold value. |
Example of the parameters
object:
Code Block |
---|
"config": { "parameters": { "maxQueryRange": 86400000, "commission": 0.1, "tolerance": 0.1 } } |
...
Code | Description | Example(s) |
---|---|---|
| This condition applies to rule processors that rely on the current transaction being successful in order for the rule to produce a meaningful result. Unsuccessful transactions are often not processed to spare system resources or because the unsuccessful transaction means that the rule processor is unable to function as designed. |
|
| For certain rules, a specific minimum number of historical transactions are required for the rule processor to produce an effective result. This exit condition will be reported if the minimum number of historical records cannot be retrieved in the rule processor. |
|
| Currently unused. | |
| The statistical analyses employed in some rule processors evaluate trends in behavior over a number of transactions over a period of time. While the trend itself can be categorized and reported by the regular rule results, some results are not part of an automatable scaled result. This exception provides an outcome when the historical period does not show a clear trend, but the most recent period shows an upturn. |
|
| Similar to |
|
Example of the exitConditions
object:
Code Block |
---|
"config": { "exitConditions": [ { "subRuleRef": ".x00", "outcome": false, "reason": "Unsuccessful transaction" }, { "subRuleRef": ".x01", "outcome": false, "reason": "Insufficient transaction history" } ] |
...
Attribute | Description |
---|---|
| This attribute defines the specific value that will be matched in the rule processor ( Every case contains a value, with the exception of the default “else” case. Values can be either strings, encapsulated in quotes, or numbers, without quotes. |
| Every rule processor is capable of reporting a number of different outcomes, but only a single outcome from the complete set is ultimately delivered to the typology processor. Each outcome is defined by a unique sub-rule reference identifier to differentiate the delivered outcome from the others and also to allow the typology processor to apply a unique weighting to that specific outcome. We have elected to assign a numeric sequence to the sub-rule references for result bandscases, prefaced with a dot (“.”) separator, but this format is not mandatory for the sub-rule reference string. Any descriptive and unique string would be an acceptable sub-rule reference. By convention, the default “else” outcome has a sub-rule reference of |
| The configuration file defines whether the result delivered by the rule processor is flagged as either |
| The reason provides a human-readable description of the result that accompanies the rule result to the eventual over-all evaluation result. |
...
id
identifies the specific typology processor and version related to its version that will be used by the configuration. There will typically only be a single typology processor active in the platform at a time, but it is possible and conceivable that multiple typology processors and/or versions can co-exist simultaneously. It is recommended that the typology processor “name” is drawn from the source-code repository where the typology processor code resides, and the version should match the semantical version of the source code as defined in the source code.cfg
is the unique version of the typology configuration. Though unlikely, multiple different versions of a typology configuration can co-exist simultaneously in the platform. The configuration consists of two parts: an arbitrary identifier for the typology to differentiate one typology from another, and then, separated by an@
, a semantical version that defines the specific version of the configuration for that typology, for exampletypology-001@1.0.0
.desc
offers a readable description of the typology
The combination of the id
and cfg
strings forms a unique identifier for every rule configuration and is sometimes compiled into a database key, though this is not essential: the database enforces the uniqueness of any configuration to make sure that a specific version of a configuration can never be over-written.
The Rules object
The rules
object is an array that contains an element for every possible outcome for each of the rule results that can be received from the rule processors in scope for the typology.
...
Every. Possible. Outcome.
All the possible outcomes from the rule processors are encapsulated in each rule’s configuration, with the exception of the .err
outcome that is not listed in the rule configuration for reasons mentioned above. When composing the typology configuration, the user must remember to include the .err
outcome, but the rest of the rule results (exit conditions and banded/cased results) can be directly reconciled with the elements in the rules
object.
Each rule result element in the rules array contains the same attributes:
...
Attribute
...
Description
...
id
...
The rule processor that was used to determine the rule result is uniquely identified by this identifier attribute.
...
cfg
...
The configuration version attribute specifies the unique version of the rule configuration that was used by the processor to determine this result.
...
ref
...
Every rule processor is capable of reporting a number of different outcomes, but only a single outcome from the complete set is ultimately delivered to the typology processor. Each unique outcome is defined by a unique sub-rule reference identifier to differentiate the delivered outcome from the others
The unique combination of id
, cfg
and ref
attributes references a unique outcome from each rule processor and allows the typology processor to apply a unique weighting to that specific outcome.
...
true
...
The outcome of the rule result will be either true or false, depending if the configurer expected the result to deterministic or not. If the outcome is true, the rule result will be assigned the weighting associated with the true
attribute in the configuration. By convention, deterministic (true) outcomes are assigned a positive number as a weighting.
...
false
...
Info |
---|
Why does the typology configuration A rule processor (defined by its id) is closely paired with its configuration (defined by the cfg): the configuration works for that rule processor and no other, and the rule processor won’t work with another rule processor’s configuration. A typology processor is a generic “engine” processor. It is not paired with a specific typology the way a rule processor is - it is intended to work for multiple, if not all, typologies. The typology configuration needs another way to reference the specific typology that will be scored by the typology processor. For that reason, the |
The Rules object
The rules
object is an array that contains an element for every possible outcome for each of the rule results that can be received from the rule processors in scope for the typology.
...
Every. Possible. Outcome.
All the possible outcomes from the rule processors are encapsulated in each rule’s configuration, with the exception of the .err
outcome that is not listed in the rule configuration because the conditions and descriptions are built into the rule processor itself. When composing the typology configuration, the user must remember to include the .err
outcome, but the rest of the rule results (exit conditions and banded/cased results) can be directly reconciled with the elements in the rules
object.
Each rule result element in the rules array contains the same attributes:
Attribute | Description |
---|---|
| The rule processor that was used to determine the rule result is uniquely identified by this identifier attribute. |
| The configuration version attribute specifies the unique version of the rule configuration that was used by the processor to determine this result. |
| Every rule processor is capable of reporting a number of different outcomes, but only a single outcome from the complete set is ultimately delivered to the typology processor. Each unique outcome is defined by a unique sub-rule reference identifier to differentiate the delivered outcome from the others The unique combination of |
| The outcome of the rule result will be either true or false, depending if the configurer expected the result to deterministic or not. If the outcome is true, the rule result will be assigned the weighting associated with the |
| The outcome of the rule result will be either true or false, depending if the configurer expected the result to deterministic or not. If the outcome is false, the rule result will be assigned the weighting associated with the |
Info |
---|
What does “every possible outcome” mean? A rule processor must always produce a result, and only ever a single result. The rule result will always fall into one of the following categories: error, exit or band/case. Results across all the classes are mutually exclusive and there can be only one result regardless of the category. Results are uniquely identified via the
The rule processor must produce one of these results (identified by the result’s |
Because the rules
object contains every possible rule result outcome from each of the rule processors allocated to the typology, the typology configuration can become quite verbose, but here’s a short example of a rules object for a typology that contains two rules:
...
Configuration documents in Actio EKUTA are strictly structured JSON documents. Each document contains an identifier related to the specific processor and version of that processor to which the configuration is to be applied. For example, the configuration for a rule processor would have the following attribute and value in the typology configuration:
...
The configuration version attribute defines the specific version of the configuration file when it is used by a processor.
Actio EKUTA employs semantic versioning3 for both processor source control and configuration documents:
...
Beyond this constraint imposed by the database, configuration versions are expected to be managed outside the platform. The Actio platform EKUTA does not currently offer a native user interface for configuration management, though Sybrin, one of the FRMS Centre of Excellence’s System Integrator partners, have created a user interface that allows for the creation of configuration documents as well as the automated management of configuration versions between iterations of a configuration document.
...
The network map defines the routing of an incoming transaction to all rules and typologies that are required to evaluate the transaction7. By default, the platform is configured to evaluate a pacs.002 transaction that concludes a transaction initiated from a pain.001 or pacs.008 message with a status response.
...
In its current configuration, the platform only evaluates the pacs.002 as the trigger payload for the rule processors and typologies have only been defined with the final status of a payment transaction in mind.
The typology processor is not currently configured to interdict the transaction when the threshold is breached; only investigations are commissioned once the evaluation of all the typologies are complete.
https://frmscoesemver.atlassian.net/wiki/spaces/FRMS/pages/76906497/Configuration+management#References%3Aorg/
An explicit version reference has been planned for development to make it easier for an operator to link an evaluation result to the specific originating network map.
We have found during our performance testing that the text-based descriptions in our processor results undermines the performance gains we achieved with our ProtoBuff implementation. We will be removing the unabridged reason and processor descriptions from the configuration documents in favor of shorter look-up codes that will then also be used to introduce regionalized/language-specific descriptions.
In its default deployment, the platform contains a single version of the “core” platform processors (the typology processor and TADProc) at a time. Though it is possible to deploy and maintain multiple parallel versions of these processors and manage routing to these processors through the network map, this guide will only focus on singular core processors for now.
Before our implementation of NATS, EKUTA processors were implemented as RESTful micro-services. The
host
attributes in the network map contained the URL where the processors could be addressed. With our initial implementation of NATS, the routing information was moved into environment variables that were read into the processors when they were deployed, or restarted in the event of a processor failure. At some point we will revert the routing information back to the network map so that we can adjust routing more dynamically while processors are in flight.