InnoDB wite activity is too high.

Hello Everybody.

I suspect that something weird is happening in my system (Percona 5.5 on CentOS 6).
Disk write activity is constantly much higher than expected.
Amazon Cloudwatch Write Bandwidth up to 8Mb/s (expected 3Mb/s)
Amazon Cloudwatch volume Write Throughput up to 300 Ops/s (expected 100 Ops/s)
Innodb_pages_written value up to 200 pages/s (expected 65 pages/s)

How do I know what to expect? I have separate system to compare with the same hardware (AWS instances) similar number of concurrent users and user activity.
We’re using replication so I know that generated bin-logs size is the same, DB growth is the same in terms of both records and size.

Could you recommend some key parameters I need to check to compare configurations of two systems which can cause this problem?

Try to increase innodb log file size ( http://www.mysqlperformanceblog.com/2008/11/21/how-to-calculate-a-good-innodb-log-file-size/ ) and increase the buffer pool size in order to reduce the number of flushed pages.

Two log files 256M each, I think this must be enough.