Download
FileDAG Storage A distributed storage service built on the IPFS technology stack. Different from the official implementation of IPFS, we focus more on data management, data reliability, availability and fault tolerance, and clustering of storage nodes. The minimum storage unit of FileDAG Storage is dag, which is the data block. Files or objects are organized in a merkle-dag structure, and multiple files or objects may share some data blocks. The obvious benefit of this is to reduce redundant data, especially for multi-version systems. Not only does reduce data redundancy, it also saves bandwidth on network transmissions. Every advantage has its disadvantage, the downside is that data management has become more complex. First, the file or object management module needs to be abstracted on the basis of merkle-dag; second, the file cannot be deleted directly, only the data blocks that are no longer needed can be released through garbage collection.
Features
- Learn from object storage services about how to do data management
- Handle the release of data blocks in a reference counting manner
- Provide data availability by using distributed storage nodes
- The underlying storage module, responsible for the storage and release of data blocks, similar to blockstore
- Object storage service abstraction layer based on Dag Pool, responsible for partially providing s3-compatible api interfaces
- Visual management UI
Project Samples
