the complexity of bringing together streaming, datalake, and schema management is holding your team back.
wings is your open-source streaming platform. it uses zero-copy iceberg storage to eliminate data duplication, giving you the power to own your entire data stack and simplify your architecture.
query engine

open table catalog

in-memory processing

storage format

wings stores topics into an apache iceberg catalog. stream data with wings, or query using any apache iceberg reader.
partition data based on your analytical use case, not your operational needs. partition data based on your schema on any key you choose, allowing for efficient querying and analysis.
define the arrow schema for your topics, wings takes care of enforcing the schema on ingestion. evolve schemas as your business needs change without the cost of migrating existing data.
wings is open source with an apache 2.0 license. you can self-host it on your own infrastructure, so your data and metadata never leaves your network.
we separate data from metadata, and storage from compute. we leverage cloud-native services for storage of data and metadata to simplify operations.
wings uses arrow flight for efficient data transfer. all data and metadata is a sql query away thanks to arrow flight sql.
wings is designed to support multiple tenants on a single cluster. projects can use different object storage backends. topics can write live data to express buckets, and use cheaper standard storage for compacted data.
we leverage liquidcache to implement distributed pushdown caching and reduce the number of object store requests, improving latency and reducing cost.
wings interacts with external services to provide authentication, authorization, and secret management. object store configuration is safely stored in external vaults.
wings is implemented in rust, leveraging the arrow and datafusion ecosystem. we provide a typescript sdk for best-in-class developer experience.
We use cookies to enhance your browsing experience.