Block Sizes, Claims and Proofs in the Multi-Gateway Era

[Olshansky (Daniel Olshansky)]

@RawthiL Firstly, ty so much for the detailed response and so promptly!

Next Steps

Below are what I’d suggest as next steps based on the data and discussion available..

Actionable now:

1. Gateways: Gateways should re-evaluate how their gigastakes are distributed and used. See the Gateway Q&A.
2. DAO: We re-evaluate what an effective pocketcore/MinimumNumberOfProofs that helps the network but does not have a considerate impact on the economics (i.e. these will become un-claimable sessions).

Education:

3. PNF Education: PNF should take this into consideration as more gateway stakes are distributed (to whom, how much, educating the gateway, etc…). Cc @Dermot

Backup bandaids:

These are last resort backups in case of emergency in order of least impact on network economics/tokenomics.

4. Increase pocketcore/BlockByteSize
   • Though LeanPOKT makes this less costly, it should not be the immediate go to option for scalability reasons.
5. Decrease pocketcore/SessionNodeCount
   • Though this will work, it will cost some decentralization and optionality w.r.t gateways managing QoS.
6. Increase pos/BlocksPerSession
   • Though this will work, the long 15 minute block time we already having, slowing it down even more is not preferable.

Follow-up Questions for @BenVan

@BenVan I have a couple of questions I wanted to follow up on.

1. You mentioned you changed something in your nodes to make sure the network has less bloat. Do you mind sharing & reiterating what that is?

2. @BenVan what are the current ssd storage requirements for a single pocket node for you? It could be a good data point to decide how viable the block size option is.

3. You suggested increasing pocketcore/MinimumNumberOfProofs to something in the low 10s.
   3.1 Am I remembering correctly?
   3.2 My understanding from the code is that NumOfProofs is literally the number of relays, so like @RawthiL said, it has to be in the hundreds to have substantial impact.

Gateway Q&A

Specific data we need to understand w.r.t gateways are:

1. How many app stakes are available to each gateway?
2. How many of the available stakes are used?
3. How are they used w.r.t to stake concentration?
4. Is there an opportunity to consolidate stake.

I’ll work on this with the backend team for Grove.

@poktblade Could you look into this on behalf of nodies?

Shannon

Relay Mining (with the relay threshold) will enable a single session to hold billions/trillions of relays by modulating the difficulty.

The Probabilistic Proof document we put together (so not all claims need a proof) will solve this part.

Nits / Questions for @RawthiL

1. Add a unit whenever the x-axis is amount so it’s more self-descriptive
2. Found a couple small typos that spell-check should catch. E.g. gruops
3. pocketcore/SessionNodeCount is mentioned twice
4. Do you know why this link shows 80MB for the block size: https://poktscan.com/explore?tab=blocks

Context

@BenVan posted the following in #node-chat.

And during the ecosystem call, we decided to collect the following data to understand next steps: