This blog was first published on the Filecoin Station blog.

This roadmap details the proposed work on the Spark Protocol in the second half of 2024. This follows conversations with the Boost, IPNI and Curio teams amongst others.

Disclaimer

By publishing the roadmap here, we are making a commitment of a medium-to-high integrity to delivering the items. However, we add the disclaimer that things can change very quickly across a bunch of different dependencies and so we cannot be certain we can deliver it all on time.

Alternative Views

You can view the roadmap in Starmap here. The Github issue that powers the Starmap view is this one. It groups the work items into five themes.

High Level Themes

• 🫂 Allocator support: the datacap process now consists of allocators who allocate their datacap to clients to make storage deals. Spark must now support these allocator “pathways” as well as the legacy FIL+ LDN pathway.
• 🎯 Data quality improvements: we continue to analyse the veracity of the data created by Spark. Spark is a trustless decentralised network and so we need to keep on top of potential attack vectors.
• 🤝 DDO compatibility: Filecoin deals are now made with direct data onboarding (DDO) as well as through the f05 Storage Market actor. Spark needs to be compatible with DDO deals.
• 🛠️ Tooling & data access: Storage providers, allocators, clients and others have asked for better tooling and access to Spark data to inform their own decisions and operations. We will work on improving the offering here.
• 👛 Incentives: Spark doesn’t work without the funds that make the Station Operators make requests. We will continue to work on more sustainable and longer term funding options for Spark, both for the client and the Storage Providers.

Roadmap

For more information about any of the work items, please click through to Github to read the description.

July 2024

• ✅ Allocator support: Support v5 Allocator pathways (Github)
• ✅ Data quality improvements: Retrieve CIDs directly using fetch not Lassie (Github)
• ✅ Data quality improvements: Report Spark retrieval errors more accurately ****(Github)
• ✅ Data quality improvements: Deterministic randomness for each SPARK round (Github)