faster production performance
Production servers are no longer burdened by cube processing, supporting faster data entry and retrieval.
darwin.Cloud performance
fresh analytics, off production
darwin.Cloud uses Kafka events and an Always On SQL Replica to update data cubes away from production workloads—keeping day-to-day activity fast while maintaining timely, accurate analytics.
darwin.Cloud cube update flow
production change
INSERT · UPDATE · DELETE
Kafka event
asynchronous queue
SQL replica
read-only source
cube update
off production
analytics
reports · spotlights
the challenge: resource-intensive updates
The data cubes behind darwin.Cloud reports, spotlights, and analytics need to update whenever data in the darwin database changes.
production server load
Running cube updates directly on production added processing work to the same servers supporting active users.
slower daily tasks
Resource competition could slow the data entry and retrieval work users perform throughout the day.
Kafka + SQL Replica
a decoupled path from data change to analytics
The redesigned process distributes work across the infrastructure while protecting the completeness of each cube update.
- 01
offload cube processing
Cube update tasks move off the production server so data entry and retrieval can remain fast and uninterrupted.
- 02
process change events with Kafka
Kafka asynchronously handles INSERT, UPDATE, and DELETE events, decoupling cube updates from active production operations.
- 03
read from the SQL replica
Updates retrieve data from an Always On SQL Replica—a read-only copy of production—rather than running heavy update queries against the primary database.
- 04
allow time for synchronization
A deliberate delay gives the replica time to synchronize before events are processed. Replicas normally update in milliseconds; a 10-minute safeguard covers unusual asynchronous delays.
why the change matters
reliable analytics without compromising daily performance
The new approach keeps analytical data current while protecting the production systems users depend on throughout the day.
efficient resource utilization
Using the SQL Replica distributes read and processing workloads more effectively across the infrastructure.
scalability
Kafka can handle a high volume of data changes asynchronously without slowing production work.
accurate, timely analytics
The synchronization delay is negligible for most use cases and keeps analytics relevant and reliable in “real-enough” time.
follow the data into the product
Connect the update process to the infrastructure and analytical experiences it supports.
data cube architecture
See how dedicated analytics servers and NVMe storage isolate cube workloads.
view detailsanalytic dashboards
Explore the brokerage KPIs and spotlights updated through this architecture.
view detailsreport samples
Open real darwin.Cloud financial, production, franchise, and agent reports.
view details