The state of Gossamer March

[P1sar]

The state of Gossamer March

Highlights

This cycle, we focused on enhancing the Candidate Backing subsystem to support elastic scaling and improve overall system efficiency. Key updates included modifications to the logic for processing messages related to candidate validation and seconding. We also completed work on Async Backing and began work on Elastic Scaling.

As part of the JAM initiative, we’ve been busy improving PVM (Parallel Virtual Machine) execution and instance management.

This week, we’re also warmly welcoming Karan Dev Sapra to the team! He’ll be contributing to the JAM implementation.

The Gossamer team, as part of the JAM Implementors DAO, is proud to join the Decentralized Voices program. Real decentralization means that builders help govern the systems they create.

Development

Polkadot Storage

Timothy continued working on the integration of the GRANDPA package into the existing Gossamer codebase.

Diego continued integrating the refactored db.Backend into our codebase.

Polkadot ELVES Implementation

Axay completed the Async Backing updates for the Candidate Backing subsystem and began work on a new candidate descriptor for Elastic Scaling.

Kiril started work on the PVF pre-checking subsystem design, which performs pre-checks on parachain runtime upgrades. This helps prevent potentially malicious runtimes from leaving nodes in a limbo state due to excessive memory usage or execution time.

Freddy started working on Availability Bitfield Distribution. Validators vote on the availability of a backed candidate by issuing signed bitfields, where each bit corresponds to a candidate. These bitfields help compactly determine which backed candidates are available, based on a 2/3+ quorum.

Haiko began work on the Availability Distribution subsystem, which handles the distribution of availability data to peers. This includes chunks, PoVs, and AvailableData (PoV + PersistedValidationData), using request-response protocols.

JAM Protocol Implementation

Jimmy completed work on the Service STF, service state serialization, accumulation state serialization, and the Reporting State Transition logic.

Eclesio and Jimmy worked on improving the PVM instance manager and initialization process.

Karan began contributing to JAM by adding missing host functions to our PVM implementation.