Hacker News Reader

Comment by grogenaut | Hacker News Reader

Dynamodb the product != the Dynamo paper which Cassy is based off. I was using Cassy in 2011 before Amazon asked me to use big bird which was the code name for DynamoDB.

View on HN

I struggle to think of any similarities, actually. I think DynamoDB's design makes it clear that the era of decentralized architecture/eventual consistency at Amazon has come and gone. DHTs just don't make sense in the data center.

Can you elaborate on this?

Do you mean that amazon is tending towards strong or externalizable consistency now?

I’ve always noticed AWS products tend towards eventual consistency, and Google Cloud offerings are almost all strongly consistent.

For example S3 is eventually consistent for read after write (if it’s not the first write). Google cloud storage is strongly consistent for read after all writes.

I think the goal is for S3 to "eventually" be strongly consistent. I heard one of S3's designers say that they shouldn't have opted for eventual consistency in retrospect.

Not sure what you mean, DDB provides both strong and eventual consistency, now in 2 fashions with transactions. 3 if you include cross region replication.

There's definitely still a DHT sitting under there.

No, it's something like a hash-partitioned primary-backup-replicated distributed database. S3 is much closer to a DHT.

There's no real difference between a DHT (distributed hash table) and a hash-partitioned distributed database. It's the same concept. Apache Cassandra is considered an implementation of a DHT.

DynamoDB is decentralized same with S3, it's just abstracted from the user.

Reply on news.ycombinator.com