May continued Akave's rapid engineering momentum with major advancements across both Akave O3 and the Akave protocol stack. This month focused on bringing O3 even closer to full AWS S3 compatibility while improving storage performance, enterprise operability, and internal maintainability. Alongside these improvements, the protocol introduced a new generation of streaming download capabilities, more efficient erasure-coded data reconstruction, flexible encryption workflows, and faster large-scale uploads.
On the O3 side, we expanded Object Lock support, strengthened versioned object handling, introduced AWS-compatible GetObjectAttributes responses, enhanced ACL and conditional request behavior, improved IAM authentication, and modernized internal repository architecture. Lower-level storage performance also improved through io_uring-based network pooling, while deployment workflows became simpler with Docker, setup, and CI enhancements.
On the protocol side, Download2 evolved into a true streaming download pipeline capable of ranged retrievals and incremental decryption. MultiUpload gained per-file encryption support and significantly improved duplicate handling, while additional optimizations reduced blockchain interactions, improved erasure reconstruction efficiency, and modernized the SDK around Go 1.26.2.
Together, these releases continue strengthening Akave as an enterprise-ready decentralized object storage platform that delivers AWS-compatible APIs while leveraging decentralized infrastructure for scalable, verifiable storage.
About Akave O3
Akave O3 is the S3-compatible object storage layer built on top of the Akave decentralized protocol stack. Every release continues to focus on three core principles:
“Security” – “Performance” – “Operability”
May's release significantly expanded all three by improving AWS compatibility, strengthening enterprise storage features, optimizing lower-level storage operations, and simplifying deployment and maintenance.
Security and AWS Compatibility
Enterprise Object Lock
One of the primary engineering efforts during May was completing additional AWS Object Lock compatibility.
Governance bypass support was introduced alongside improvements covering object retention, multipart uploads, CopyObject operations, versioned deletes, and delete-marker handling. Numerous edge cases identified during QA testing were resolved, ensuring behavior more closely aligns with AWS S3 across complex object lifecycle operations.
Version-aware object management also became significantly more reliable through corrections to latest-version tracking, delete handling, cache invalidation, and repository consistency.
AWS-Compatible S3 Behavior
O3 continues closing the remaining compatibility gaps for applications built against AWS S3.
This release added:
- GetObjectAttributes support with AWS-compatible responses
- STANDARD storage class normalization
- Conditional request handling
- Checksum propagation across S3 operations
- Improved CopyObject support for versioned objects
- AWS-compatible ACL grant parsing
- Better object tagging compliance
- Correct handling of versioning responses and delete markers
These improvements reduce behavioral differences between O3 and AWS S3, allowing existing applications to migrate with fewer modifications.
IAM Improvements
The IAM subsystem continued evolving with both architectural improvements and stronger authentication behavior.
Key additions include:
- O3IAM_ access-key prefixes
- Access-key prefix routing for authentication
- Improved entity organization
- Collection indexing
- Expanded repository validation
- Additional authentication test coverage
These changes improve authentication flexibility while preparing IAM for future enterprise capabilities.
Performance and Runtime Efficiency
Storage and Network Optimization
Performance improvements extended into the storage engine with the introduction of io_uring-based pooled network operations.
This new architecture improves concurrency during asynchronous file and network operations while reducing overhead for high-throughput storage workloads.
Additional improvements included:
- Multipart upload metadata helpers
- Improved checksum processing
- UploadPartCopy race-condition fixes
- Better backend initialization
- Improved storage copy utilities
Together these changes increase reliability while providing better scalability under heavier workloads.
Repository and Metadata Improvements
Large portions of the storage repository were reorganized throughout the release.
Repository logic was separated into focused implementations for individual operations including bucket management, object listing, validation, copy, retrieval, and deletion. Object listing also received expanded filtering, validation, and repository-level optimizations.
Expanded automated testing now covers many more repository edge cases, improving long-term maintainability while reducing regression risk.
HTTP and Runtime Stability
The HTTP layer also received several reliability improvements.
Server initialization was decoupled to improve concurrency safety while keep-alive handling, address formatting, and connection behavior were refined. HTTP test infrastructure was modernized to eliminate brittle assertions and provide more stable continuous integration results.
Enterprise Deployment and Operability
Deployment workflows continued becoming simpler for operators managing Akave clusters.
Docker images were streamlined through shared builder base images, reducing maintenance while improving build consistency.
Setup improvements introduced:
- Better support for sync-object volume mounts
- Improved checksum synchronization
- Enhanced Traefik integration
- Direct S3 deployment improvements
- More flexible IAM and Console configurations
Continuous integration pipelines also received substantial modernization through dependency caching, simplified package installation, improved workflow organization, and expanded documentation across repository, storage, HTTP, metadata, and AIO components.
Protocol Layer Upgrades
Akave Protocol v0.5.6
The protocol stack focused on making decentralized downloads faster, more memory efficient, and easier to integrate into production applications.
The largest addition this month was the continued evolution of Download2 into a true streaming transfer architecture.
Streaming Download2
Rather than reconstructing an entire encrypted file before writing it to disk, Download2 now performs streaming reconstruction with incremental decryption.
This dramatically reduces memory usage while allowing applications to begin receiving data immediately, making the protocol much better suited for large AI datasets, media files, and enterprise workloads.
Ranged Downloads
Download2 also introduced native ranged download support.
Applications can now retrieve only selected byte ranges from encrypted erasure-coded files without downloading the complete object. Supporting APIs now expose both original and encoded file sizes required to accurately reconstruct requested ranges.
This functionality enables far more efficient large-scale data access while reducing unnecessary network traffic.
MultiUpload Improvements
MultiUpload became significantly more flexible through several major improvements.
Per-file encryption keys are now supported, allowing every uploaded file to maintain its own encryption configuration. Encryption settings also moved into the IPC API, simplifying SDK configuration while providing more granular control.
Duplicate handling was substantially improved as well. Rather than failing an entire batch upload when duplicate files are encountered, individual files now report their own errors while allowing the remainder of the upload to continue successfully.
SDK Performance Improvements
Several improvements focused on reducing both blockchain interactions and transfer overhead.
Batch uploads now fetch bucket information only once and reuse that data throughout the upload process, significantly reducing chain calls during MultiUpload operations.
Developers can also configure stripe sizes and encrypted block sizes through SDK options, providing greater flexibility while maintaining proper alignment validation for streaming erasure-coded transfers.
Faster Reconstruction
Download performance also improved through smarter erasure reconstruction.
Instead of downloading every available erasure block, reconstruction now selects sixteen random blocks while maintaining recovery guarantees. Combined with new in-place erasure decoding, this reduces network traffic, memory allocations, and overall download latency.
Platform Modernization
The protocol development environment also received several foundational improvements.
These include:
- Go 1.26.2 adoption
- Updated project dependencies
- Regenerated protobuf bindings
- Better SDK test organization
- Expanded multinode testing
- Updated Docker images
- Improved CI workflows
These updates strengthen the development platform while preparing future protocol releases.
Release Recap
O3 v1.7.0
A major AWS compatibility and enterprise storage release.
Key highlights:
- Expanded Object Lock compatibility
- AWS-compatible GetObjectAttributes
- Improved versioning and delete-marker behavior
- Enhanced IAM authentication
- io_uring network pooling
- Repository architecture refactoring
- Better Docker and deployment workflows
- HTTP reliability improvements
Akave Protocol v0.5.6
A major streaming transfer and SDK optimization release.
Key highlights:
- Streaming Download2
- Native ranged downloads
- Per-file encryption keys
- Improved MultiUpload
- Faster erasure reconstruction
- Reduced blockchain interactions
- Go 1.26.2 upgrade
- SDK modernization
What's Next
Streaming-first data transfer remains one of the primary engineering priorities across the Akave platform.
Future work will continue expanding the Download2 and Upload2 architecture while further improving AWS compatibility, enterprise management capabilities, and storage performance.
Upcoming areas of focus include:
- Continued AWS S3 compatibility improvements
- Production hardening of streaming transfers
- Additional storage performance optimizations
- Expanded enterprise IAM capabilities
- Improved observability and operational tooling
- Larger-scale throughput tuning
Connect with Us
Akave Cloud is an enterprise-grade, distributed and scalable object storage platform designed for AI, analytics, decentralized infrastructure, and large-scale enterprise workloads.
It provides:
- S3 compatibility
- Cryptographic verifiability
- Immutable auditability
- SDKs for agentic AI systems
- Zero egress fees
- No vendor lock-in
- Up to 80% lower storage costs compared to hyperscalers
Akave works with a broad ecosystem of infrastructure providers operating across multiple regions and countries, enabling sovereign and distributed data infrastructure at scale.