Status update for a work done in September

[P1sar]

Status update for a work done in September

Optimizing the polkadot-sdk dependency footprint

• Continue work on fixing issue regarding non-direct dependencies that brings curve25519-dalek dependencie with different versions. W've forked verifiable and bumped schnorrkel (which pulls in curve25519-dalek-ng in older versions), but building polkadot-sdk against that causes weird errors related to serde. Also bumped version of the schnorrkel library
• Added CI check for unused dependencies
• Added duplicate-deps check in CI

WebRTC

We have finished adding libp2p WebRTC server and client to litep2p performance tests
Progress was delayed due to WebRTC-direct being effectively non-functional in litep2p, caused by a defect in the str0m dependency with DTLS fingerprint verification that we also fixed.
The same is currently almost done for smoldot as a client and litep2p as a server. No PR yet, but here is a documentation on work done this direciton.
After finishing transport layer performance testing between litep2p and smoldot we will start integrating them on the applicaton layer.

Speculative availability

We have finished all the unit-tests, becnhmarking and currently finishing zombinet tests, unfortunatly for last couple weeks we were stuck with some unexpected behaviour during zombinet tests that we are trying to debug and fix.

Approval rewards:

Since the last update we have adjusted our implementation plan.

As a first step, we have implemented a new subsystem that is responsible for collecting all new statistics regarding validator rewards.

The next step will be to include this subsystem in the most recent release in order to collect important metrics. These metrics will help us understand how rewards might influence economic security and token distribution for validators.

While metrics are being collected, the following step will be to complete RFC-119, since it currently does not define implementation details about reward statistic collection and distribution.

Current research proposes to use fully on-chain statistics submission. This would require every validator to submit their subjective view about rewards that should be distributed once per epoch. This approach will require approximately 1024 (validators) * 12 bytes (~12 KB) of storage usage per epoch, but will reduce network usage and make the implementation easier to build and debug.

Smart Contracts Infrastructure Integration (Foundry-Polkadot)

This month, we initiated work on the Anvil forking functionality. At this stage, a PoC has been developed, which is capable of forking certain simple networks in a way of regenesising local node, hence it is far from necessary design. Forking of asset-hub is not yet supported, primarily due to the lack of inherent transaction mocking. The PoC serves as an initial exploration to assess the complexity of the implementation.

In subsequent iterations, the roadmap includes the introduction of lazy loading, comprehensive inherent transaction mocking, and a dynamic host function trait implementation to improve flexibility and maintainability.

Off-chain runtime

Work onthis feature is postopned until Approval Rewards task is moved to testing phase.

DDoS protection

Work on this task is postponed due to work on Polkadot-Foundry