-
Scaling near to 1M TPS
We achieved 1M TPS on NEAR after many optimizations targeted at increasing throughput and scaling up the number of shards, only using commodity hardware. Below is a detailed look at the techniques that made it possible.
-
Random Sampling of NEAR’s Validators
NEAR’s current approach is sufficient for the current number of shards and validators, but it can easily become inefficient if the number of validators grows. This blog post presents a new, simple and efficient way of selecting NEAR’s validators and analyzes its safety and liveness guarantees.
-
Enhancing NEAR Tokenomics: Version 2.9.0 Halving Upgrade
To support a more sustainable tokenomic model and strengthen incentive alignment across the ecosystem, nearcore version 2.9.0 reduces the maximum annual inflation rate of NEAR Protocol from 5% to 2.5%. This upgrade will become effective once 80% of stake of block producing validators choose to adopt it. For more details on upgrading your mainnet node, please see the 2.9.0 release notes.
-
Mainnet Decreased Block Production Rate
NEAR mainnet briefly slowed on deployment of new smart contracts after the validator set grew past 256: the broadcast routine that erasure‑codes contract byte‑code for all validators relied on a Reed–Solomon library hard‑capped at 256 recipients, so the chunk producer that included a
DeployContracttransaction crashed during the final “code distribution” step, though the finished chunk still propagated. The Near One team released a hot‑fix that simply truncated the recipient list, which worked for any nodes that upgraded to adopt the fix, but the fix didn’t account for the creation of a Reed-Solomon encoder on un-upgraded nodes, which caused any un‑upgraded node to crash when it tried to forward those contract parts, temporarily enlarging the blast radius; block throughput dipped to ~0.1 blocks/s for about half an hour and then recovered as most validators adopted the patch.
subscribe via RSS