description |
---|
Datafast Runtime is a high-performance subgraph processing runtime which is written from scratch and designed to handle subgraphs with unparalleled speed & storage-efficiency |
Code repository at https://github.com/vutran1710/datafast-runtime
Document repository at https://github.com/vutran1710/datafast-runtime-document
❤️ Performance-focus - it is built to prioritize indexing speed
❤️ Blockchain data reusability - block data is stored in Object Storage for reusability, eliminating cost of calling JSON-RPC over and over again for the same data
❤️ Blockchain data is smartly organized & compressed using Databricks' DeltaLake technology plus Protobuf on top for storage-efficiency, very fast serialization, very fast to ingest and unlimited scalability
❤️ Indexed data is stored in NoSQL database to reduce database cost yet still fast in retrieval
❤️ Allowing user to specify a limit time to keep output data in store, like last 6 months or last year - help reduce data storage cost by remove outdated data
❤️ Graceful reorg-handling
❤️ Extensible - as Datafast code base is way simpler and extensible than graph-node, allowing user to freely implement more features without crying for help
❤️ No need IPFS!
❤️ No GAS required for indexing!
❌ Datafast Runtime is built to focus on the indexing-job. It does not come with a GraphQL query layer.
Datafast is built for educational & sharing purpose & not ready for production - since there are plenty of features that needs implementing & issues need addressing.
Still, Datafast-Runtime can be used as a ground-work to build in-house blockchain indexing solution to replace GraphNode / TheGraph
- Companies who want to host their own subgraph data
- Companies who want to host their own subgraph data with efficient storage cost, with their own solution
- Developers who want to play around with their own version of Subgraph Indexing Engine
- Developers who want to play around with WebAssembly technology and build crazier stuffs with it.