VRF Resilience and Stability Update
We launched a new stability-improving upgrade for Randomizer today after experiencing a short delay in callbacks last Saturday.
In the realm of decentralized networks, Randomizer has been a shining beacon of reliability. Our Verifiable Random Function (VRF) protocol, currently hosted on Arbitrum One mainnet, has consistently maintained near 100% uptime since its deployment 50 days ago. However, last Saturday (May 13th) at approximately 3AM GMT the protocol experienced an unforeseen delay in VRF callback fulfillments for a few hours, which we want to transparently address and clarify. We also promptly launched a new solution that prevents any such disruptions from happening in the future.
🔍 Understanding the Delay
The disruption was caused by an unusual event, a block finality issue on Ethereum. This edge case event resulted in the decentralized beacon system having connectivity issues when it came to posting VRF callbacks.
Beacons, the cornerstone of our VRF protocol, rely on callbacks to maintain their registered status. As the issues unfolded, the inability to post callbacks resulted in beacons becoming unregistered over time. Eventually, the majority of the beacons were unregistered as a result, leading to the Randomizer protocol being unable to make callbacks to VRF requests for a few hours.
Despite this disruption, it's essential to note that no VRF callbacks actually failed. The incident merely introduced a delay in the execution of these operations.
All VRF callbacks were eventually completed, preserving the integrity and security of the Randomizer protocol.
🏋️♂️ Ensuring Future Resilience
We understand that any disruption, no matter how temporary or minor, can affect your contract’s users' experience. Therefore, we have promptly launched an update to prevent disruptions like this from happening again.
After a beacon is unregistered, it must be manually re-registered (currently by the Randomizer contract owner, later to be decentralized). Due to a previous lack of internal notifications systems for developers and beacon operators, the beacons were re-registered at a delay.
We swiftly took action and introduced a new solution yesterday, within 24 hours of discovering the issue: a new system that immediately notifies beacons of any disruption to their operations and lets them re-register themselves quickly and efficiently (with the click of a button).
This system empowers beacon operators with the ability to quickly review, repair, and, if necessary, re-register their beacons themselves.
Moreover, to enhance the resilience of our protocol, Randomizer will also be equipped with automated AI-powered repair and re-registration capabilities, where a language learning model performs automated tasks to review, repair, and re-register a beacon. In the event of any future disruption, beacons will quickly be able to restore and continue to provide reliable service to all users.
🤝 Commitment to Reliability
These measures are a testament to our unwavering commitment to maintaining the high reliability of the Randomizer VRF protocol. By enabling rapid beacon self-repair and self-registration, we aim to ensure that Randomizer maintains its 99.99% uptime and real-time results, regardless of any disruptions that may occur on the underlying network.
We understand the trust you place in us, and we are devoted to continuously improving and hardening our systems against any potential stability issues. With these implementations, we strive to make Randomizer an even more resilient and reliable VRF protocol for all developers.