We need to keep a footprint as low as possible on the machine running a Mithril signer. We have noticed that the current process to import the Cardano transactions does not allow an efficient pruning strategy on the signer: the block range roots are computed only once all the transactions are imported. The import mechanism should be able to compute block range roots along the way in order to allow early pruning and overall low disk space requirements.
What
Compute block range roots as soon as they can be computed by the signer and allow for early pruning on the transaction store.
How
[x] During warmup phase on the signer, run the import of transactions with the Transaction importer on a sequence of ranges of block numbers (e.g. every 10 block range root length or 150 blocks)
[x] Implement a decorator on the Transaction importer to import by sequence of block ranges
[x] Assess the required disk space for a signer node (~100MB with 90M transactions)
Why
We need to keep a footprint as low as possible on the machine running a Mithril signer. We have noticed that the current process to import the Cardano transactions does not allow an efficient pruning strategy on the signer: the block range roots are computed only once all the transactions are imported. The import mechanism should be able to compute block range roots along the way in order to allow early pruning and overall low disk space requirements.
What
Compute block range roots as soon as they can be computed by the signer and allow for early pruning on the transaction store.
How
During warmup phase on the signer, run the import of transactions with the Transaction importer on a sequence of ranges of block numbers (e.g. every10block range root length or150blocks)~100MBwith90Mtransactions)