The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →iceberg-catalog-rest 0.10.1 could reach three of the seven catalogs tested in the reported setup: local Apache Polaris, Google BigLake, and Microsoft OneLake. The author reports that the client could express 13 of 25 selected distinct endpoints, and that supported reads answered on all three services. AWS Glue and Amazon S3 Tables did not work through this REST client in the reported signature experiment; Databricks Unity Catalog and Snowflake Horizon were not tested. These results are specific to the versions, dates, and setup below—not a guarantee of compatibility with other releases.
What the test measured—and what it did not
The test examined Apache Iceberg’s Rust REST catalog client, iceberg-catalog-rest 0.10.1, alongside iceberg 0.10.1. It checked whether selected catalog requests could be expressed and whether requests received answers from particular services. That is narrower than proving every feature works: an answered request does not establish that the returned data is semantically correct, nor that a table’s data files can be read.
As an Amazon Associate I earn from qualifying purchases.
The author reports that the client could express 13 of 25 distinct endpoints in the selected test set. Eleven were absent and one existing method was a stub. A separate count of 19 of 33 checks counts repeated probes individually; five checks used the same update_table endpoint. The 13-of-25 figure is the more useful measure of distinct endpoint coverage, but neither count represents the whole evolving Iceberg REST API.
The report names iceberg-storage-opendal 0.10.1, reqwest 0.12.28 with rustls-tls, and rustc 1.98.1. Source was read on September 4, 2026; versions were captured September 17; managed and local catalog runs took place September 18; and a local write run took place September 21. The author used one machine in one region. Those dates and versions matter: catalog support, crate releases, and service behavior can change.
#1 Best Overall
Results across the seven catalogs
| Catalog | Evidence in the report | Reported result for the tested REST client |
|---|---|---|
| Apache Polaris | Run locally | Seven supported reads and 11 writes answered. The local server used permissive settings, and a local table-file read succeeded. |
| Google BigLake | Managed service run | Seven supported reads answered using an externally minted gcloud token. The reported GCS-backed metadata read succeeded. |
| Microsoft OneLake | Managed service run | Seven supported reads answered using an externally minted az token. The attempt to read the table metadata file timed out. |
| AWS Glue | Signature experiment only | Requests were refused in the tested setup because this REST client did not provide AWS Signature Version 4 (SigV4) signing. The project has a separate Rust Glue catalog crate. |
| Amazon S3 Tables | Signature experiment only | Requests were refused in the tested setup because this REST client did not provide SigV4 signing. The project has a separate Rust S3 Tables catalog crate. |
| Databricks Unity Catalog | Not run | No measured service result. The report’s comparison is based on source, not a test against Unity Catalog. |
| Snowflake Horizon | Not run | No measured service result. The report’s comparison is based on source, not a test against Horizon. |
The read and write counts are the author’s observations under this test configuration, not independent compatibility ratings or population-wide statistics. The AWS entries likewise describe a limited signature experiment, not a full service evaluation.
Authentication is a deciding factor
In the reported setup, the REST client handled token or header-based authentication paths. The BigLake and OneLake runs used tokens minted outside the tested client: one via gcloud and one via az. That means the results show the requests could be made with those supplied credentials; they do not show the crate itself obtained those credentials.
AWS was a different case. Glue and S3 Tables required SigV4 signing for the tested catalog requests, which the REST crate could not provide in that setup. A fixed authorization header is not a substitute for request signing: a signature is bound to a particular request and cannot simply be reused for another request. The Rust project’s separate Glue and S3 Tables catalog crates are the described alternative, but they are distinct implementations rather than a switch that adds SigV4 to the REST crate. Their Catalog operation behavior also differed from the REST client and from each other in the versions examined.
The Apache Iceberg Rust project repository lists REST, Glue, and S3 Tables components. For an actual application, check the release-specific crate documentation and issue tracker before choosing an implementation; a mutable project branch may not describe version 0.10.1 exactly. The project’s tracker described SigV4 support for the REST crate as an open enhancement at the time covered by the report.
Rank #3
Catalog access does not guarantee table-file access
A catalog resolves metadata and exposes table operations; reading a table also requires access to the storage location holding metadata and data files. The report’s successful catalog calls therefore should not be read as proof that every table could be loaded.
For storage, the tested setup used the OpenDAL-backed Iceberg storage crate. The article reports that its test path did not use catalog-issued storage credentials. This distinction is especially relevant to OneLake: catalog reads answered, but reading the table metadata file timed out. The report also describes a successful local file read and a successful GCS-backed read. It does not establish why the OneLake attempt timed out, so the result is a boundary of that test rather than proof that OneLake storage is generally unreadable.
The Apache Iceberg REST specification describes storage credentials in catalog responses and remote-signing configuration for storage providers. Specification support alone does not establish that this particular Rust client version consumes those features, and those storage mechanisms do not by themselves add SigV4 signing for AWS catalog API requests.
What to check before adopting the client
- Pin the actual crates. Confirm the version of
iceberg-catalog-restand any storage crate your application will use; the reported results concern version 0.10.1, not every Rust Iceberg release. - Match the authentication flow. Determine whether your catalog accepts credentials your application can supply, or requires signing that the chosen implementation does not perform.
- Verify the operation you need. The report covered a selected set of endpoints, not the entire REST API. Check whether your workflow depends on an endpoint absent from that set or on the existing stub.
- Test the storage path separately. After catalog operations succeed, try loading table metadata and reading the relevant files with the exact storage backend and credentials your deployment will use.
- Do not treat untested rows as compatibility claims. Unity Catalog and Horizon were not run in this report; their rows are source-based comparisons only.
HTTPS setup also deserves a release-specific check. The report’s build notes say the consuming application enabled a reqwest TLS feature for HTTPS, while an Apache Rust issue tracked TLS features for the REST crate as an open enhancement at the time. Check the crate release actually in use rather than assuming the configuration is unchanged.
Practical verdict
If your target is Polaris, BigLake, or OneLake, the reported tests provide evidence that this version of the REST client can perform the tested catalog reads with the described credentials. For Glue or S3 Tables, the report points to the corresponding dedicated Rust catalog crates rather than the REST client. For Unity Catalog or Horizon, this report cannot answer whether the client works. In every case, validate the storage access and the specific operations your application needs in your own version and environment.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
