Rockset adds MySQL connector to real-time indexing database
Rockset made a set of connectors for MySQL and PostgreSQL generally available. The tools enable real-time data indexing for analytics on transactional data.
Real-time indexing database vendor Rockset advanced its vision of a converged database with the general availability of new integrations with the MySQL and PostgreSQL databases.
One key challenge, however, has been with operational data stores where data is stored in a relational database such as MySQL or PostgreSQL as a system of record. Rockset is aiming to overcome that problem with a direct integration that makes it easier for data from MySQL- and PostgreSQL-compatible databases to be replicated into Rockset, where users can then use the data for real-time analytics.
Rockset made the new integrations generally available on July 8. The MySQL and PostgreSQL integration connectors have been available as a preview technology since April and have had a few early users.
Igor Blumberg, technology director at Dimona, said the company uses Amazon Aurora, a MySQL-compatible cloud database. Blumberg noted that Dimona's database system controls and keeps data on all aspects of the company's operations including finance, inventory, production and sales.
Rockset for inventory tracking
A problem Dimona has faced is how to keep track of its inventory accurately and in real time.
Blumberg said the company maintains several warehouses. Within each warehouse are multiple locations for inventory, such as shelves or pallets that contain clothing items. Blumberg referred to these locations as addresses that are constantly shifting over the course of a day. Meanwhile, Dimona needs to keep track of where all its inventory is.
Rockset provides a set of data connectors that help pull in data from a source destination so that it can be indexed in the Rockset real-time database.
To help enable real-time data visibility, Dimona replicates the Aurora database to Rockset and then queries Rockset for the current inventory. Blumberg said the live replication of MySQL to Rockset has made it easier keep all the data updated.
"We don't need to worry about keeping the database cache layer updated," Blumberg said. "We can focus on our business and let Rockset calculate the inventory."
How Rockset's MySQL and PostgreSQL integration works
Rockset co-founder and CEO Venkat Venkataramani explained that users had previously also been able to pull in data from MySQL and PostgreSQL databases.
We don't need to worry about keeping the database cache layer updated. We can focus on our business and let Rockset calculate the inventory.
Igor BlumbergTechnology director, Dimona
Venkataramani said that the difference between the new MySQL and PostgreSQL integration and what users were doing before is that the previous process was more manual and slower. Rockset users could pull data in a batch mode at periodic intervals such as every hour or every day. In contrast, the new integration enables a real-time data replication.
Rockset first scans and copies all the data from an existing MySQL or PostgreSQL database. After that initial data copy, Rockset moves to what Venkataramani called a "fast follow" model.
That model provides real-time data replication by using change data capture (CDC) capabilities that are already present in both MySQL and PostgreSQL. CDC provides a stream of all new data inserts, updates and deletions to existing records in a database.
Venkataramani explained that as data streams in from the source MySQL or PostgreSQL database via CDC, it is automatically indexed in Rockset and turned into fast SQL query tables.
By enabling users to pull in data from an online transaction processing (OLTP) database such as MySQL and PostgreSQL, Venkataramani said that users could now also benefit from the online analytical processing (OLAP) capabilities of Rockset's real-time indexing database.
Venkataramani said that the vendor's roadmap is to continue to build out integrations and connectors for data that can be pulled into the Rockset database.
"No matter where your data lives, we want to build a real-time connector to that data set," he said.
_mobile.jpg)
