Explore

5X Platform: Architecture of the fully 5X hosted offering

In the fully 5X cloud hosted offering - both the control plane and the data plane of the platform is deployed in 5X’s cloud in a region closest to the customer.

⁠

⁠

Question

Explanation

How is data being consumed from our internal sources (e.g., RDS)?

A lightweight 5X‑managed connector container (Airbyte‑compatible) pulls the authorised rows/columns over a secure private link and streams them into the Kubernetes cluster.

How is data being processed (ETL) on your side?

Ingestion pipelines and transformation models run in a k8s cluster. Additionally, other platform workloads including modeling, data apps, semantic layer, BI also run in an siolated k8s cluster.

How is data (temporarily) stored on your side (caching, etc.)?

Rows that are ingested as part of ETL pipelines exist only in ephemeral pod volumes for the duration of a job. We persist only metadata (job status, column stats, lineage) in the Control Plane’s Metadata DB.

How do you feed data back into our warehouse (e.g., Redshift)?

The paired destination connector writes batch files or streaming inserts directly into your Redshift cluster or S3 bucket; once the write succeeds, the pipeline deletes its working data.

There are no rows in this table

⁠

Key Architecture Highlights

Clear isolation

Control Plane – Authentication, platform UI/APIs, Metadata DB, Billing Engine.

Data Plane – Regional Kubernetes cluster running 5X‑Managed Connectors, dbt Core, BI services, data‑app runtimes.

Regional deployment: Entire 5X stack is instantiated in the customer’s chosen AWS/GCP/Azure region; data never leaves residency.

Ingestion at scale: 600 + connectors out of the box that are deployed in a k8s clusters with auto‑scaling for high‑throughput CDC and SaaS APIs alike.

Ephemeral processing, no retention: After the destination write is acknowledged, connector pods delete local fragments; only metadata persists for audit and billing.

⁠

Want to print your doc?
This is not the way.

Try clicking the ⋯ next to your doc name or using a keyboard shortcut (

CtrlP

) instead.