Volante Technologies chief technology officer Krishna Sudarshan espouses a federated approach to reference data in a recently issued what paper, Reference Data – an Insight into This Thorny Problem. Sudarshan promotes the idea of using Volante’s Message Metadata Repository and other technologies to build and maintain central metadata rules without the need for a single physical data store.
According to Sudarshan, because of its pervasive nature, reference data requires that any technology solution for its management is comprehensive in its coverage. “Any solution,” he says, “has to broadly address the following four problem facets: Data acquisition, transformation, storage/analysis and deployment.”
It is also essential that downstream or internal applications can be easily integrated with the reference data solution. He also adds that the addition of new vendors, data sources and instruments means that no solution can be a one-off effort: the solution needs to be maintained properly.
Sudarshan notes that firms acquire data from a multitude of sources, each typically delivered using the provider’s proprietary formats requiring proprietary application programming interfaces (APIs). Volante uses so-called Format Plug-ins to hide message idiosyncrasies and allow users to focus on the message content rather than formatting details.
For example, Volante’s Swift format plug in allows users to send or receive corporate actions data over the Swift network. It also has a built-in data dictionary that allows users to construct new messages based on the ISO 15022 standard.
Whatever its source, some data may not need to be saved into a firm’s data model. According to Sudarshan, “some data may have to be simply transformed into a firm’s internal format and directed towards any applications that may utilize this data.” This is particularly true in the case of market data, for example, which needs to be handled expeditiously, he says.
Rather than having each application perform these transformations separately and each time there is a change to a source format, Sudarshan believes a simpler approach is to convert data into a common standardized or ‘canonical’ format irrespective of its ultimate destination. XML, he says, is increasingly becoming the format of choice as an internal or standardized format.
A second type of transformation deals with the issue of correlations, he says. “For instance, transformations are needed to standardize entity identifiers… A firm may standardize securities identifiers internally using correlation schemes which look up the incoming identifier and replace it with a standard identifier that is used within the organization.”
He adds: “If the canonical format is based on a data dictionary, interpretation and analysis of the incoming data is unambiguous.” He points out that a range of pre-built data dictionaries are available today, including ISO 15022, MDDL and FpML.
One possible approach to the challenges of reference data management, Sudarshan says, is encapsulated in Volante’s federated approach. “Volante’s Message Metadata Repository centralizes metadata rules related to data element definitions, message definitions and transformations,” he says. “However, groups within a firm can choose to deploy these rules on different platforms based on the individual needs. Centralization of rules provides a firm all the benefits of data centralization without the associated constraints of a physical data store.”
The metadata repository includes the firm’s data dictionary and message catalogue. The messages in the message catalogue are comprised of business elements in the data dictionary. Applications within the firm may have different message formats, with the result that they may not have been designed for optimal interoperability. Internal messages in the repository therefore may have to be transformed to and from an application’s proprietary format. Volante captures such transformations as message models, or Cartridges, which may feature all transformations pertaining to a particular application or vendor feed.
Using this kind of approach, Sudarshan says, “firms can not only implement an initial solution, but also continue to maintain this solution to address new and impending regulatory and business demands imposed upon it.”
