Tech Talk: Cassandra Data Modeling

In this talk, Patrick McFadin (Chief Evangelist for Apache Cassandra, DataStax) breaks down topics like storing objects, indexing for fast retrieval, and the application life cycle.

This talk was given at Cassandra Day Silicon Valley 2014.

ABOUT DATA COUNCIL:

FOLLOW DATA COUNCIL:

[patrickmcfadin]

Right. The remaining parts of the primary key specify how the data is stored in the storage row.

[fdk]

Why use MD5 for hashing, isn't it quite expensive for the purpose? Also, how do you expand the token ring? I guess you have to repartition the entire database unless you exactly double the nodes.

[ramans]

16:52 Are placement (Node in the cluster) and the Partition Key i.e. the Storage Row within the Node, both determined by the first part of the primary key?

[nareshgb]

Th epoint about covered indexes in oracle is this: It will STILL do random reads, because "logically consecutive index blocks" are not "physically consecutive blocks on disks". But of course, there are SSDs.

[nareshgb]

"single seek on disk" - do we really use "disks" these days? or is it a "bunch of disks with striping and other memory layers" (aka storage servers) ? or is it SSDs?

[jgoemat]

If you create a clustered index on all three columns then all the data is stored in the index and you won't need random reads...

[dgrader]

It is clear that the guy giving the talk learnt his share of technology from just talks ( people talking to him and him talking to people ). The whole talk lacks depth on the topic. First 20 mins of the talk he is still talking about data indexing which is common in RDBMS.

[detoxorg]

Terrible sound recording killed this talk for me.