You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Storage Providers should be able to serve requests to the IPFS network as several storage clients, such as web3.storage, Estuary, and IPFS pinning services, and their end users are part of the IPFS network. Adding support for Bitswap removes the need for intermediary services, such as autoretrieve. Leveraging the flexibility of Boost we should enable this as a standalone service, similar to booster-http, so that it can be independently scaled as needed.
As part of this effort we are working with the Kubo and IPFS.io gateway teams to enable e2e tracing of the retrieval pipeline as part of our efforts to improve retrieval reliability. This will ensure we have visibility into each component in the retrieval pipeline to better resolve reliability and performance issues. The metrics/logging added for Boost will also enable us to troubleshoot retrieval problems seen by other retrieval transports (Graphsync and HTTP) at lower layers of the Boost code base (DAG store, piece store, etc)
Milestones
1. E2E Integration with Kubo for free, public content
With the completion of this milestone Kubo will be able to query Network Indexers, such as cid.contact, for a given CID and discover SP's who are running the Bitswap transport.
Requirements
SP's must be able to restrict access to content via existing retrieval access patterns
Only 0 cost retrieval will initially be supported
Bitswap content records should be discoverable on network indexers, such as cid.contact
Summary
Storage Providers should be able to serve requests to the IPFS network as several storage clients, such as web3.storage, Estuary, and IPFS pinning services, and their end users are part of the IPFS network. Adding support for Bitswap removes the need for intermediary services, such as autoretrieve. Leveraging the flexibility of Boost we should enable this as a standalone service, similar to
booster-http
, so that it can be independently scaled as needed.As part of this effort we are working with the Kubo and IPFS.io gateway teams to enable e2e tracing of the retrieval pipeline as part of our efforts to improve retrieval reliability. This will ensure we have visibility into each component in the retrieval pipeline to better resolve reliability and performance issues. The metrics/logging added for Boost will also enable us to troubleshoot retrieval problems seen by other retrieval transports (Graphsync and HTTP) at lower layers of the Boost code base (DAG store, piece store, etc)
Milestones
1. E2E Integration with Kubo for free, public content
With the completion of this milestone Kubo will be able to query Network Indexers, such as cid.contact, for a given CID and discover SP's who are running the Bitswap transport.
Requirements
Issue tracking
Enabling the Bitswap retrieval server
Adding metrics/logging
Related Issues
References & Related Items
The text was updated successfully, but these errors were encountered: