Note - info for this section taken from RDC Pricing Architecture powerpoint in \\thenest\Product Management\RDC Pricing\Project_Administration\Presentations
The ultimate goal of RDC Prices is to produce Gold Copy Price records for a given effective date and a fixed set of securities out of raw prices received from multiple vendors, or by using Prior Day Gold Prices. To achieve this goal, RDC Prices uses Data Strategies, which process through the below stages:
RDC Prices uses Data Strategies to produce Gold Copy Price records for a given Effective Date and set of securities. Data Strategies encapsulate the necessary conditions to satisfy and actions to take, by processing through the below stages:
Security selection
Source Hierarchy
Demand phase
Validation phase
Compositing and Release phases
A more detailed explanation of these stages is below:
Data Strategies
Data Strategies process activity by working through different phases.
Security Selection
Data Strategies maintain their own security criteria.
Data Strategies can reference one or more security holdings (Portfolios), each with their own selection criteria.
The final set of securities subject to processing by a Pricing Data Strategy is determined by first applying the Data Strategy Security Criteria against the Securities database, then applying all Holdings security criteria to the resulting set.
Source Hierarchy
Data Strategies reference a Source Hierarchy where all participating vendor sources are arranged in order of importance.
The Prior Day Gold Copy source is optional, and usually assigned the lowest level in the hierarchy.
Demand Phase
The processes that load vendor data into the database need to know which prices are expected for a specific Data Strategy.
The Demand Phase is executed in order to create “shell” records for each combination of security, source from the source hierarchy, and effective date.
The Demand Phase also creates a Gold Copy “shell” record for each combination of security and effective date.
Validation Phase
Data Strategies use Validation Rules to validate source prices during the Validation Phase.
Each Validation Rule has its own Selection Security Criteria that is used against the set of securities that was used to generate demand.
If a vendor price record fails validation, then it cannot be used later in Gold Price compositing.
Validation rules can be global or regular:
Global rules are applied to every Data Strategy using global validation rules
Alternatively, Data Strategies can declare use of specific regular validation rules while ignoring global rules
Compositing and Release Phases
For each security, the direct outcome of the Validation Phase is the Best Vendor Source Price record
Best Vendor Source Price record is chosen based on validation results and source hierarchy rules.
During the Compositing and Release Phases, the Best Price record is converted into the Gold Price record
Data Strategies can use specified Enrichment Rules that are applied to the Best Price record as part of the compositing process.
PACE Server Engines
Engine is basic execution thread used by PACE servers to perform a specific task concurrently with others.
Each engine has independent access to the database and other resources.
RDC Prices uses the concepts of a Main Engine and Rule Engines.
Main Engine vs Rule Engine
The Main Engine is always invoked first
The Main Engine interprets the security criteria in a Data Strategy to produce a list of security aliases subject to processing
The Main Engine determines which Data Strategy phases to be executed and launches Rule Engines for each phase using a predetermined order.
Rule Engines are responsible for execution of specific phases against a subset of security aliases
The Main Engine can be executed either synchronously or asynchronously, while Rule Engines are always executed asynchronously.
The number of Rule Engines required for each phase is determined by the following factors:
Maximum number of securities processed by a single Rule Engine is limited through system configuration
One Rule Engine can only process one Validation Rule per Data Strategy.
Add label
0 Comments