Google introduces the PostgreSQL AlloyDB database

At the Google I/O event, Google showcased AlloyDB for PostgreSQL, a fully managed, PostgreSQL-compatible database for demanding enterprise-level transactional and analytical workloads. Google claimed that in its performance tests, AlloyDB delivered “more than 4x faster on transactional workloads and up to 100x more analytical queries than standard PostgreSQL, all with simple and predictable pricing.”

Google advises in a blog post that AlloyDB for PostgreSQL was built on the principle of compute and storage disaggregation and designed to take advantage of disaggregation at every layer of the stack.

Sign up for your weekly dose of what’s happening in emerging technologies.

Google explains that AlloyDB starts by separating the database layer from storage, then introduces a new intelligent storage service optimized for PostgreSQL. Google claims this reduces I/O bottlenecks. The storage service itself also disaggregates compute and storage, allowing block storage to scale separately from log processing.

Design

It is a distributed system and has three parts:

  • A low-latency regional log storage service that enables very fast log write-ahead.
  • A log processing service that processes these write-ahead log records and produces “materialized” database blocks.
  • Regional partitioned, fault-tolerant block storage for durability even in the event of zonal storage failures.

Image: Google

Google has listed some key advantages of this approach. These include full compute/storage disaggregation, storage layer replication, efficient I/O paths/no full page writes, low latency WAL write, rapid creation of read replica instances, fast reboot recovery, and storage layer backups.

Maria H. Underwood