Data retention is one of my favorite database topics because it has one of the greatest impacts on database performance. It also has a great impact on the capabilities of a company to fully utilize information for profit and efficeiency.
Keeping too much data causes you database performance to degrade, increases the length of maintenance windows, and can even result in the inability to perform effective disaster recovery or failover strategies. Increased data also results in increased cost for storage as more storage devices are required.
There are lots of strategies for reducing the cost of data retention.
- Copy and/or summarize data and store it in a data warehouse.
- Place data on different data devices. New, or active data, is stored on high performance drives backed up frequently. Older data is stored on lower performing devices and does not require backing up as it is static. Sometimes there may be a combination of the two. Other implementations take advantage of SSD for the most highly used data.
- Copy data into another database with slower, less expensive disks
- Backup and delete stale data
- Start fresh databases annually
You may not have a data retention strategy. One thing that is important is to not find you need one in the middle of a crisis. Don’t wait until your application performance has declined to unacceptable levels, and you don’t have enough hardware resources to begin implementation because you can’t add any more load to resolve the issue.
What’s your opinion? Is a data retention policy, or implementation important? Maybe it doesn’t matter for everyone. If not, then who needs to consider data retention techniques? Share your thought here or by email to btaylor@sswug.org.
Cheers,
Ben