Back Issues/Search Home → Calendar → Archive → RSS → Subscribe → Services → Vendors → Blog → Current Issue → Popular → Newsletters (LinkedIn, Substack)

All issues › Volume 344, Issue 2 › IT Vendor News › Databricks

Dispelling LTAP Myths

Databricks, Thursday, November 12th, 2026

Databricks addresses common misreadings of Lakebase's LTAP design, in which the lake itself is the Postgres database's storage.

Databricks responds to commentary on Lakebase's LTAP architecture, which keeps one durable copy of each table in the lake as a Delta or Iceberg table under one catalog.

It rejects the claims that a single copy forces either point lookups or scans to be slow, that lake tables are inherently stale, and that Postgres logical replication already solves extraction.

The post distinguishes replication, which keeps two databases with separate catalogs and timelines aligned, from tiering, which puts caches in front of one database.

In Lakebase, safekeepers, pageservers and compute caches are tiers in front of object storage, with point lookups reading Postgres heap pages and scans reading Parquet.

more →  ·  More from Databricks →