Mars is a small but perfectly formed “microservice”, that allows you to present it with two records and have it index, standardise and compare them together so you can understand how similar they are in real life scenarios.
Mars doesn’t use traditional relational or deterministic technology, in fact it uses a rules extended probabilistic engine called “Identiza” that was developed by Robert Haynes over the past decade. This type of engine is often called “Fuzzy” Matching and has been the basis for many traditional MDM products, however the problems with these technologies are that they are incredibly hard to understand and also very complex to extend.
EntityStream Mars is a much simple way to process data, instead of relying on converting all data sources into one form so they can be compared and contrasted, Mars instead allows you to store the data in a closer to nature form, that doesn’t require any significant processing before it can be matched. Match rules and comparisons are done between heterogenous data structures and this means the chances of error is greatly reduced due to the lack of loss of data quality when you transform data.