Performance Tuning Recommendations

This section describes some recommended performance tuning configurations to optimize BAM. It assumes that you have set up the BAM server on Unix/Linux, which is recommended for a production deployment. If you have high volume of data with high concurrency, we also recommend you to use a distributed BAM setup. For instructions, see our product deployment and clustering guide.

The values we discuss below are general recommendations only. They might not be the optimal for the specific hardware configurations in your environment. We recommend you to carry out load tests on your environment to tune the BAM accordingly.

Improvement Area Performance Recommendations

Data receiver nodes

Xms1024m -Xmx1024m -XX:MaxPermSize=512m
Change the /etc/security/limits.conf with the following values:
```
* soft nofile 4096
* hard nofile 65535
```

Analyzer nodes Xms1024m -Xmx1024m -XX:MaxPermSize=512m

Dashboard nodes Xms1024m -Xmx1024m -XX:MaxPermSize=512m

Hadoop nodes

Recommended OS: Linux
At least 10 GB storage capacity in each node
At least 100 Mbps network bandwidth
Set hadoop.root.logger=ERROR

Cassandra nodes

Have separate disks for your commit log and data dirs (sstables)
Set the Heap memory as follows:
System memory Heap size
Less than 2 GB 1/2 of system memory
2-4 GB 1 GB
More than 4 GB 1/4 system memory, but not more than 8GB
Set following configuration in cassandra.yaml according to your hardware resources:
- concurrent_reads: 4 * no of cores
- concurrent_writes: 8 * no of CPU cores

Also see http://www.datastax.com/docs/1.0/operations/tuning in Cassandra documentation.