Hacker News Reader

S3 support is built into Apache Hadoop [1]. Azure Blob Storage & OpenStack Swift are also included (see Hadoop Compatible File Systems section in the left sidebar). A properly provisioned NameNode can support hundreds of millions of blocks. Like any high performance system, it won't do this out of the box and does require some tuning based on your use case.

[1] http://hadoop.apache.org/docs/current/hadoop-aws/tools/hadoo...

View on HN

> Like any high performance system, it won't do this out of the box and does require some tuning based on your use case.

Absolutely, I'm just trying to discern whether it can reasonably be considered a "second choice" after S3 today.

I think what I'm finding out here is that the choice is highly dependent on the workload. I suppose there isn't anything terribly surprising in that.

Reply on news.ycombinator.com