Ultimate Go: Advanced Engineering Episode 5

Cheikh Seck

December 16, 2022

Introduction

In case you missed episode 4, Bill attempted to integrate current database sharing solutions to solve the “database ownership” problem his dependency manager faces. He tried providing copies of the database, implementing a load balancer and replication. While performing these experiments, Bill identified the qualities his dependency manager’s database needs to have for users to trust it. He closed episode 4 by stating that his dependency manager needs a database that is distributed, cryptographically auditable, synchronized and scalable. By implementing this Bill hoped to provide greater transparency for his dependency manager.

In this video, Bill will borrow a few concepts from blockchain to ensure his dependency manager meets the end user’s technical requirements. Bill will define a simple proof-of-work consensus (PoW) algorithm and provide a step-by-step for each design decision he made along the way. He will use this algorithm to ensure atomic writes are performed on the database, and in turn, maintain a “single” database. Although ideal, Bill will point out the disadvantages of this algorithm as well. Follow along and learn how data is written to a blockchain and how it manages to have identical data across different nodes.

Things you will learn in this video

• Atomic write refers to the database handling transactions one at a time.
• Consensus algorithms can handle “atomic” appends.
• Proof-of-work (PoW) algorithm can be summarized as the process of applying brute force calculations to solve a mathematical puzzle and the first node to finish wins the chance to write to the database.
• PoW is inefficient, resulting in the creation of proof-of-stake (PoS) and proof-of-authority (PoA) algorithms.

Video