The Accord Consensus Protocol, how to do ACID transactions at scale

Conference

Room: Linux

Scheduled at : Thursday 11:15 12:15

Abstract

The Accord Consensus Protocol, providing global leaderless single-network-round-trip consensus using commodity clocks.

Research from University of Michigan & Apple Inc. introduces ACID-compliant, strict serialisable transactions possible globally at any scale, at high throughput, with low latency.

Move over Paxos, ePaxos, Raft, Janus, Calvin, all now outdated consensus protocols, slow and fault intolerant.

We'll explain why your database must be leaderless, scalable & fast. The notion of copying data to analytics platforms is dead weight in the age of Data Mesh & Soft-AI. Your database can no longer be a single-process monolith from ages decades old if you are to be David against the Goliath of Big Tech.

Mick will explain hard problems in computer science solving elastic scalable ACID-compliant low-latency databases, explaining trade-offs from sharding data, shadow paging, sagas, to choosing eventual consistency & isolation consistency from the database.

Mick Semb Wever

Mick Semb Wever

A 11-year Cassandra veteran, a member of the Cassandra Project Management Committee (PMC), this year's PMC Chair, and an Apache Member.

Bi-cultural with homes in both Australia and Norway.

Currently working at DataStax, as a Principal Architect for Open Source. Our Open Source solutions: Apache Cassandra, Apache Pulsar, K8ssandra, Stargate, and Kaskada, provides real-time data platforms for AI-Driven Applications to the world's biggest clusters and global enterprises.

I love engineering and coding, but nothing beats being helping others become successful.

Speaker's bio
Presentation type Conference, 45mn
Track Big Data, Machine Learning, IA & Analytics
Presentation level Expert and senior
Keywords isolation database Distributed Systems Big Data Event-Driven Microservices

Room Linux