Marcus Eagan, Product Manager, MongoDB, took to the stage and Unqork Create 2022 to present MongoDB Atlas Search, which delivers application search experiences 30% to 50% faster and simplifies management by combining three systems—database, search engine, and sync mechanisms—into one.
“You all know the power of search. You don't need to know a query language. You don't need to know SQL or MQL or whatever newfangled query syntax comes up,” said Eagan. “And that's why many of the most valuable products and companies are all powered by search.”
Yet most developers rely on brute-force “like” or “contains” queries. They are easy to implement but are very limited in their capabilities. And they can slow down developers because they return results too slowly and/or too many results that are irrelevant.
Like our brains
“We needed to find a way that would enable people to find information more akin to how their brains find information,” said Eagan.
To solve that problem, MongoDB’s solution uses a data format known as dense vectors, which works in an associative way like the brain, training models based on occurrence and co-occurrence of words.
MongoDB’s solution also simplifies the sync pipeline ETL. You gain a single, unified API across both your database and search operations, simplifying queries and reducing development time.
“We've gotten rid of having the concept of the transactional database, the search cluster to support the searches, and the sync pipeline between those two systems, which previously probably involved at least three drivers for interacting with data systems,” explained Eagan. “Now there's one driver, so you reduce the number of dependencies, there's one cluster, and there's one API.”
In addition to searches that are 30% to 50% faster, MongoDB Atlas Search means there is no need to stand up and manage a sync mechanism, write custom transformation logic, or remap search indexes as your database evolves.
Learn more about MongoDB Atlas Search.