Your data is the backbone of your company’s productivity. Your database stores every aspect of your business, so you can never be too careful when creating a backup plan. The more time between backups, the more data you lose after a disaster. Disaster recovery plans help reduce the duration of downtime after a disaster, but the key to a successful full recovery is a good backup plan. A backup plan along with a disaster recovery plan can save you potentially millions of dollars in data loss from fire, flood, theft, cybersecurity event, or any other disaster that could hit your business.
Backup Frequency Depends on your Tolerance for Data Loss
The initial thought that goes through most business owners’ heads is that they should create a full backup as frequently as possible, but full backups are expensive to host. You need the storage space and bandwidth to store these files. A large database’s full backup could potentially require terabytes of storage space. It’s standard practice to have a month’s worth of data backups, so this option can be expensive and unnecessary.
A better way to determine backup frequency is to consider your loss tolerance. Loss tolerance is determined by calculating how much data can be lost before the organization loses too much revenue. It’s also determined by calculating the amount of time an organization can be down before incurring too much revenue loss.
Loss tolerance is dependent on the organization. Usually, a large corporation performs several database transactions every hour. For large global enterprises, it can be as frequently as several transactions a minute. This means that the loss of data can be much more expensive for organizations that lose data over an hour versus smaller ones that only perform a few transactions within the same timeframe. The first question you should ask yourself is how much revenue loss would be incurred should the organization lose an hour of data. If it’s too much, you need more frequent backups.
What Types of Backups Do You Need?
Every backup schedule starts with a full backup. A full backup is a copy of all database objects and data. This backup will be as large as your database, including the transaction log. The transaction log is a record of every change on the database, so you should back up this component as well.
After a full backup is completed, you now need to decide between differential or incremental backups. Most organizations work with incremental backups, but differential backups are optional for faster restore times. Another option is to work mainly with incremental backups but create the occasional differential backup to make data restoration easier.
Incremental backups contain only the data that has changed from the last backup. This backup file will be much smaller than a full backup, because only data that has changed will be stored. The last backup could be a full backup or another incremental backup. When you restore data using an incremental backup, you must have all incremental backups executed after its related full backup. Most organizations use incremental backups until a specified time and then create a new full backup to reduce the number of incremental backups that must be stored.
A differential backup takes a snapshot of data that changed after the full backup was created. This means that if your last backup was two weeks ago, the differential backup file will contain all data that changed from two weeks previously until the current date regardless if you’ve created any other differential or incremental backups. These backup files are much larger than incrementals, but they will allow you to restore data faster since you need fewer files during disaster recovery.
Even if you use incremental and differential backups, you will eventually need to start over again with a full backup. The length of time you work with incremental backups depends on your preferences, but it’s standard to work with incremental backups for at least a week before you create a full backup. For larger organizations with many more transactions, a full backup could be created at the beginning of each day, and then incremental snapshots are taken every few minutes or every hour.
Where Do You Store Backups?
Backups should be stored in a secure location. This could be in the cloud or on the local network. A good disaster recovery plan for backups follows the 3-2-1 rule. The rule says:
1. You should have at least three copies of your backup files.
2. At least two copies of your backups should be stored on two independent media. For example, you could store backups in the cloud and on the local network.
3. One copy of your backups should be located off-site.
By using the 3-2-1 rule, you ensure that all backups don’t fail when a disaster strikes. For instance, if you store backups in the cloud and on a network drive, you will not lose your data after a fire incident that burns the local copy. The 3-2-1 rule is good for business continuity after a disaster, and it will save you time when you need to recover data at a temporary site until damages are repaired at your local office location.
Cloud storage also saves you money. It’s much cheaper to provision resources in the cloud, because you only pay for the storage space, bandwidth and computing power that you use. With local storage space, you must purchase expensive equipment, pay for someone to install it, and then pay personnel for its maintenance. With cloud storage, you provision directly from a dashboard, and the cloud provider performs all maintenance.
Conclusion
No matter what the size of your business, you need a database backup to ensure that you can recover from a disaster. The frequency of the backup schedule depends on your loss tolerance and the speed at which you need to restore data before your company suffers from catastrophic revenue loss. The right backup plan enhances your disaster recovery plan, and it can ensure that you don’t lose any data that might be critical to your business productivity and continuity.
