Delta Lake 3.3.3 fixed Delta Sharing deletion vector caching issues and upgraded client authentication.

Delta Lake 3.3.3 introduced important stability corrections for Delta Sharing consumers by resolving a deletion vector cache bug that previously caused long-running reads to fail [5]. The maintenance release also upgraded the embedded Delta Sharing client library, which improved OAuth token management and connection retry behavior during high-volume remote reads [5]. In parallel, Delta Lake 4.4.0 expanded Unity Catalog Delta API integration to Delta Kernel and the experimental Apache Flink connector, broadening external read and write access for catalog-managed tables [4].

Alongside core engine updates, Databricks product leadership presented OpenSharing across several community sessions, detailing an architectural path that builds on Delta Sharing protocols for distributed AI data exchange [3]. These technical discussions focused on how multi-region architectures and global data distribution patterns reduce cloud egress expenses while improving throughput for remote consumers [1].

Adopters testing these patterns also uncovered early operational boundaries in mixed-format environments. Practitioners reported API failures when attempting to execute the OpenSharing Get Table endpoint against foreign Apache Iceberg tables registered within their catalog [2]. The report highlighted practical interoperability hurdles that data engineering teams encountered when applying OpenSharing workflows to non-Delta table formats [2].