Seaplane Managed Global SQL (MGSQL) is Seaplane's global relational database offering based on Postgres 11. MGSQL functions like any other Postgres database, meaning it is fully wire compatible with Postgres 11.2, but with some powerful additions.
MGSQL provides you with a singular entry point
sql.cplane.cloud that routes
to the nearest available cluster relative to where you are in the world. You can learn
more about connecting to MGSQL here.
Powerful Features for Global Users
MGSQL extends standard Postgres with powerful features designed for global users and end users. Like all Seaplane products, MGSQL deploys your database globally on the best combination of edge, bare metal, and public cloud resources. MGSQL even examines your end-user traffic patterns to determine the best deployment locations based on cost and performance.
However not all data can or should be accessible everywhere, so MGSQL also allows you to restrict data locations by table or by row to comply with data privacy regulations. Deployments will still optimize your infrastructure mix for you, but only within the geofenced locations you define. You can learn more about geofencing, applying business rules, and MGSQL's various table types here.
Multi-Region and Multi-Cloud
By default, MGSQL deploys your database globally across clouds and edge, automatically picking and choosing resources and locations to minimize latency for your end users. MGSQL currently supports over 35 locations which you can read about here
Global Database, World-Class Performance
"Global" might imply "slow" (there is a lot of ground to cover) but MGSQL does not compromise on performance. We tested the database with the following queries using Sysbench from hosts operated on AWS in the EU and Japan:
INSERT INTO users (username, email, address, seaplane_geofence) VALUES ('generated-name', 'generated-email', 'address', '<country/jp or region/xe>');
-- select count
SELECT COUNT(*) FROM users WHERE seaplane_geofence='<country/jp or region/xe>';
-- select string
SELECT * FROM users WHERE username LIKE '%john%' AND seaplane_geofence='<country/jp or region/xe>';
We ran each query at least 1000 times. Each query consistently scored well below 25ms for the 95th percentile.
|query||user location location||db host location||# of queries||avg latency||95th percentile|
- Data Regulation Compliance with Seaplane SQL
- Going multi-region with Seaplane Managed Global SQL (Coming Soon!)