Reltio Connect

 View Only

How Reltio Differs from Relational DB-Based System Under the Hood

By Gino Fortunato posted 08-23-2021 11:30

  


Reltio is a revolutionary Master Data Management (MDM) system.  It is designed to make data management simpler, easier to use, and more scalable compared to traditional systems that are built using relational databases.

Reltio accomplishes these goals by reimagining how an MDM system should work under the hood.  By scrapping and simplifying the underlying data structures used by legacy MDM systems, Reltio creates a more scalable MDM solution that offers capabilities that are difficult or impossible to implement in traditional systems.

How Relational Databases Work

Relational databases are based on tables.  A table consists of a set of columns outlining the different features and a collection of records with data for each of these features.

A single table can contain a lot of data, and relational databases expand their capabilities using interrelated or linked tables.  By building complex queries involving one or more tables, a user can access any desired collection of data.

A primary limitation of relational databases is their complexity.  While each table may be relatively simple and well-structured, an organization may need tens, hundreds, or thousands of tables to contain all of the data that it requires.  Additionally, these tables are all interrelated, resulting in a complex and fragile system.  This complexity and fragility is driven by the fact that the change needs to be done in multiple places within the schema.  This requires a number of test iterations to get right.

How Reltio Works

Unlike relational databases, Reltio is not based on the concept of interrelated tables.  Under the hood, Reltio uses a much simpler structure.  In Reltio, all data is stored in two parallel lists:

  • Entities: The entity list covers ‘nouns’: grammatically any “person, place or thing” within the system, including customers, products, households, TV shows, etc.  These entities are related to other entities in a variety of different ways.
  • Relationships: The relationship list describes the relationships between entities by matching two entities in a one-to-one pairing.  For example, one relationship may indicate that a particular customer purchased a particular product.  Another relationship may state that a doctor has admitting privileges at a given hospital.

This simplified structure makes Reltio much more scalable, flexible, and usable than traditional database-driven systems.  New relationships can be added by updating metadata, which contains the details of the relationship.  Investigating an entity’s relationships involves querying this metadata rather than accessing and mapping out the relationships between dozens or hundreds of interrelated tables.

Reltio’s Unique Structure Provides Unique Capabilities

Reltio’s unique structure makes it a more compact and simple way to store and access the same data that is traditionally stored within relational databases.  However, this increased efficiency and usability is not the only benefit of Reltio’s approach.  Reltio’s unique structure makes it capable of accomplishing things that are difficult or impossible with relational databases, such as:

    • Automated Merging: A system may have multiple different records for the same entity, which should be merged into a single record.  By analyzing relationships, Reltio can perform this minimization process automatically.  For example, the discovery that Kathryn and Kate used the same credit card for different purposes indicates that these are the same person and the records should be merged.
    • Suggested Relationships: By storing all information as entities and relationships, Reltio makes it easy to detect when multiple entities both are related to the same entity.  This enables Reltio to automatically suggest relationships.  For example, the fact that Alice and Bob have the same address indicates that these two customers are likely part of the same household.
    • Survivorship: Consultants will tell you that the trickiest part of legacy MDM implementation is getting the groups in an organization to agree on a single set of survivorship rules.  Reltio enables a new survivorship strategy called Aggregation which takes all the values across all the systems.  Try putting all of that into a single cell of a relational database!  Additionally, Reltio pioneered rulesets that allow multiple groups to have their own survivorship rules giving them the contextually relevant view of the data that they need to run their part of the business.  
    • Easy Field Addition: With a relational database, adding a new column to a table is a complex process that can impact all of the interrelated tables within an MDM solution.  As a result, tables must be carefully designed and can take months to change due to its effect on merge tables, match tables, and Extract-Transform-Load (ETL) processes.  With Reltio, defining a new attribute is simple and has zero impact on other records within the MDM system.
  • Complex Merge Scenarios: As suggested above, a key feature of a master data application is the ability to merge multiple records. In a relational database, this is done by overwriting on the rows of a table, destroying that row and the ID that came with it.  Over the course of the lifetime of a record, this might happen dozens if not hundreds of times.  The destruction of the row might not be a particular problem if it were not for the fact that sometimes these merges needed to be rolled back, due to a change in business rules or due to an error in the original merge.  These complex merge scenarios are difficult to handle with relational databases but are easy for Reltio and our non destructive merge process.

Simplifying Master Data Management with Reltio

A business’s data is a core part of its ability to compete in the marketplace.  If a company’s data is siloed or is embedded in an MDM solution that is difficult to manage and use, then it isn’t providing full value to the company.

Reltio has revolutionized MDM by replacing legacy, table-based MDM with a different approach to data management.  By treating everything as an “entity” and focusing on mapping the relationships between entities, Reltio offers a more efficient and effective solution to the MDM problem and unlocks capabilities that are unavailable in traditional, table-based systems.


#Featured
#masterdatamanagement
#Blog
0 comments
3721 views

Permalink