[POC] Native Flink s3 fileSystem

[Samrat002]

What is the purpose of the change

Native flink s3 Filesystem

Why flink-s3-fs-native ?

1. Validate that Flink can talk to S3 directly via AWS SDK v2 without Hadoop, This enhancement simplifies the stack and removes dependency on Hadoop & Presto thus depreacating flink-s3-fs-hadoop and flink-s3-fs-presto and eventually remove it.
2. Reduce classpath conflicts and shading complexity caused by Hadoop + S3A; make S3 support self-contained and easier to operate.
3. Unify behaviour for cloud-first deployments (Kubernetes, containers, IAM roles, path-style/virtual-host) with straightforward configuration in flink-conf.yaml.
4. Establish a foundation to implement Flink-specific features (entropy injection, recoverable writer, retries/backoff) natively rather than inheriting Hadoop semantics.

Things that need attention

1. There is a need for a clear strategy to deprecate and remove flink-s3-fs-hadoop and flink-s3-fs-presto.
2. Users relying on Hadoop configuration/kerberos auto-wiring must map configs to native options; operational playbooks change.
3. Coexistence with flink-s3-fs-hadoop and flink-s3-fs-presto can fragment testing and user guidance if not clearly instructed.
4. S3A-specific behaviours won’t automatically carry over; must be implemented deliberately to avoid regressions.