Ultimate Go: Advanced Engineering Episode 7

Cheikh Seck

December 19, 2022

Introduction

In episode 6, Bill gives an overview of what consensus algorithms are and how these algorithms ensure distributed databases are in sync. Moving forward Bill will adopt the proof of authority (PoA) algorithm as a means to determine which node has the ability to write the next record. To jog your memory, PoA is a consensus mechanism where identity is used as a stake. By implementing this algorithm, Bill will add transparency to his dependency manager as he is no longer the only user with the ability to add new records. The dependency manager now meets the teams’ requirements for transparency, however, it will have low throughput since only one record can be shared and inserted at a time.

In this video, Bill will add a new architectural change to his database and introduce a new data structure to queue records that will be added to the database. He states that this data structure will act as a memory pool and must be synchronized across the different nodes. He sets these requirements to ensure that there is no ongoing duplicate work and nodes starved from writing have the chance to insert a new record. Watch along as Bill gives a walkthrough of how a blockchain solves the throughput problem introduced by adopting consensus algorithms.

Things you will learn in this video

• Memory pool is a batch of records to insert into a database.
• Memory pools increase efficiency as it prevents 2 nodes from trying to insert the same record.
• Why/How audits will be performed with the memory pool.
• A batch of records is a block on the blockchain.

Video