• 🔍 Problem Definition:
• 🌍 Context:
• 💣 Problem:
• 🎯 Solution:
• ⏱️ Planning:
• Team:
• Tentative plan:
• 📒 Notes

🔍 Problem Definition:

🌍 Context:

At the current stage, fast retrieval for data stored with Filecoin deals can be provided via the storage of an extra copy of the files (see 🏎️Product Problem: Fast Retrieval), but there is no actual guarantee that this extra copy is stored and maintained by the Storage Provider.

While windowPost periodically checks that SPs keep storing the replicas, we do not have an equivalent for extra copies.

This is particularly important in the context of Fil+ deals, where SPs are given a storage power multiplier for accepting deals on real data that get reviewed by notaries. Especially for these deals it is important to guarantee the possibility of quickly accessing the files upon request.

💣 Problem:

We aim to provide periodical guarantee of accessing Fil+ deals via the extra copy, in order to have a probabilistic proof of the possibility of fast retrieval.

Of course this is not a guarantee of retrieval, but we already put in place an ad hoc protocol to achieve that (see 📦Retriev.org).

🎯 Solution:

We aim to use a dedicated periodical proof, similar to WindowPost, targeted for extra copies of Fil+ deals. The hope is to use circuit similar to the WindowPost one, in order to minimize complexity.

⏱️ Planning:

Team:

• researcher: nicola, irene, luca (DRI)
• engs: ?

Tentative plan:

• Phase 1: Protocol Design (irene, luca)
• Gathering info
• Actual Protocol Design [1 week]
• Phase 2: Protocol Spec (irene, luca)
• Spec Writeup
• Spec Review
• Phase 3: implementation (nemo/jake ???)