Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Amazon DynamoDB Transactions

aws.amazon.com

Post reply on HN
Print view View original

1–10 of 77 posts

Re: Amazon DynamoDB Transactions

#2

Post by piinbinary »

My wishlist for DynamoDB is now down to:

* Fast one-time data import without permanently creating a lot of shards (important if you are restoring from a backup)

* Better visibility into what causes throttling (e.g. was it a hot shard? Was it a brief but large burst of traffic?)

* Lower p99.9 latency. It occasionally has huge latency spikes.

* Indexes of more than 2 columns

* A solution for streaming out updates that is better than dynamodb streams

Re: Amazon DynamoDB Transactions

#3

Post by jedberg »

This is cool, it lifts the burden of having to bake "atomicity" into your app if you're using a key/value store like DynamoDB. I can see a nice balance of combining this with some built in error checking in the app itself.

Re: Amazon DynamoDB Transactions

#5

Post by monstrado »

piinbinary wrote: post #2

My wishlist for DynamoDB is now down to: * Fast one-time data import without permanently creating a lot of shards (important if you are restoring from a backup) * Better visibility into what causes throttling (e.g. was it a hot shard? Was it a brief but large burst of traffic?) * Lower p99.9 latency. It occasionally has huge latency spikes. * Indexes of more than 2 columns * A solution for streaming out updates that…

Also, better insight into partition sizes / what's causing hot spotting. The DB abstracts a lot from the user, which isn't necessarily great, because it's still subject to the normal pitfalls of a NoSQL database.

Re: Amazon DynamoDB Transactions

#6

Post by jchrisa »

Congrats to the DynamoDB team for going beyond the traditional limits of NoSQL.

There is a new breed of databases that use consensus algorithms to enable global multi-region consistency. Google Spanner and FaunaDB where I work are part of this group. I didn’t catch anything about the implementation details of DynamoDB transactions in the article. If they are using a consensus approach, expect them to add multi-region consistency soon. If they are using a traditional active/active replication approach, they’ll be limited to regional replication.

Re: Amazon DynamoDB Transactions

#7

Post by darkr »

piinbinary wrote: post #2

My wishlist for DynamoDB is now down to: * Fast one-time data import without permanently creating a lot of shards (important if you are restoring from a backup) * Better visibility into what causes throttling (e.g. was it a hot shard? Was it a brief but large burst of traffic?) * Lower p99.9 latency. It occasionally has huge latency spikes. * Indexes of more than 2 columns * A solution for streaming out updates that…

Not a particularly easy solution, but you can use dynamo streams to achieve this by loading fast into a temporary table, trickle-feeding via a stream into another table. When it’s caught up, stop writes on the import table then swap over to the permanent table.

A way of doing this without expending all that effort is oh my wish list too.

Re: Amazon DynamoDB Transactions

#9

Post by erik_seaberg »

jchrisa wrote: post #6

Congrats to the DynamoDB team for going beyond the traditional limits of NoSQL. There is a new breed of databases that use consensus algorithms to enable global multi-region consistency. Google Spanner and FaunaDB where I work are part of this group. I didn’t catch anything about the implementation details of DynamoDB transactions in the article. If they are using a consensus approach, expect them to add multi-region…

They warn about other regions seeing incomplete transactions (if you opt into transactions on global tables), which fits with the current "copy each new item from the stream" async replication.
Post reply on HN