* Posts by datamgmt

2 posts • joined 4 Oct 2021

Yugabyte's double-decker DBaaS follows Cochroach in distributed RDBMS


Re: Relational?

Hi Robert

I know a bit about Yugabyte - the base design is inspired by Google Spanner, but with the added advantage of a PostgreSQL compatible interface - specifically, because it uses the PostgreSQL v11.2 query layer.

There is a great article about how an Atomic Clock can be replaced by a Hybrid Logical Clock https://www.yugabyte.com/wp-content/uploads/2021/05/Distributed-Transactions-Without-Atomic-Clocks.pdf

The added advantage of this approach is that it can be used cross-cloud and hybrid on-prem/cloud as it doesn't rely on a specific atomic clock configuration but has the same minimum transaction time guarantees.



Open-source veteran PostgreSQL emits release 14: Tweaked, scalable, and ready to get heavy


Re: Yugabyte - "distributed relational database "

Hi W.S.Gosset

Yugabyte YSQL supports both SERIALIZABLE and SNAPSHOT isolation levels (they map to the PostgreSQL isolation level syntax of SERIALIZABLE and REPEATABLE READ respectively). Snapshot Isolation level is the default for YSQL.

SNAPSHOT Isolation guarantees that all reads made in a transaction will see a consistent snapshot of the database, and the transaction itself will successfully commit only if no updates it has made conflict with any concurrent updates made by transactions that committed since that snapshot.

SERIALIZABLE Isolation guarantees that transactions run in a way equivalent to a serial (sequential) schedule.

This is powered by the underlying architecture where the core database uses a modified RocksDB with a hybrid real-time clock and MVCC to ensure transactional consistency.

I don't see retailers like the giant Krogers in the US committing themselves to Yugabyte if what you claim is true - the new generation of PostgreSQL API compatible, distributed databases are going to dominate the largest scale deployments and this great new release of PostgreSQL will consume the smaller and medium size implementations.


Biting the hand that feeds IT © 1998–2022