S3 Compatibility Protocol

we-trust-s3 implements an Amazon S3-compatible object storage interface, allowing YYKV to serve as a storage backend for massive unstructured data.

1. Architectural Implementation

Unlike traditional file systems, YYKV builds the S3 protocol layer directly on top of the HAL (Hardware Abstraction Layer):

RESTful Interface: Provides standard S3 APIs via HTTP/HTTPS.
Virtual Buckets: Maps the S3 Bucket concept to YYKV's tenant-level logical namespaces.
Object Mapping: Key is mapped to the storage engine's keys, and Body utilizes yykv-io's Direct I/O for large-block continuous writes.

2. Supported APIs

Service Operations: ListBuckets
Bucket Operations: GetBucket (List Objects), HeadBucket
Object Operations:
- PutObject: Supports Multipart Upload optimization.
- GetObject: Supports Range requests, enabling physical-level offset reads.
- DeleteObject: Atomic deletion.
- HeadObject: Retrieves metadata.

3. Storage Optimization: Large Object Slicing

For S3 scenarios, YYKV employs specific storage strategies:

Automatic Sharding: Objects larger than 4MB are automatically sliced into multiple physical Pages.
Hybrid Storage: Object metadata is stored in high-performance SSDs (Blaze tier), while object body data automatically migrates between SSDs and high-capacity HDDs based on access frequency.

4. Security & Authentication

Compatible with the standard S3 V4 signature algorithm (AWS4-HMAC-SHA256):

Clients sign requests using AccessKey and SecretKey.
YYKV's internal authentication module verifies signatures and associates them with the corresponding TenantID.

5. Use Cases