Guest Column | February 3, 2017

Backup, Restore, And Disaster Recovery

Rich Rutkowsi, Sale Point Data, LLC

By Rich Rutkowski, Sale Point Data, LLC

Before deciding on a backup technology there should be consideration made towards your business process and how it is impacted by a loss to compute capability. What follows is a list of those considerations, but first we need to distinguish between Restore and Disaster Recovery.

Restore is when you have a working computer and lost some data or a program; Disaster Recovery is when you do not have a working computer and lost all data or no longer have a computer and place which to restore it to. The Backup/Restore and Disaster Recovery process is not something we add to an existing system, rather it should be influenced by the business process and disruption should be minimized.

What And Where To Back Up
You need to determine if you are going to backup only the data that you generate, the applications, the OS, or the entire computer system. This is basically driven by how long you can afford to spend restoring a system and how much backup storage will be required. Also decide if you want to protect against lost data, a disaster, or both. Simply backing up your data can protect against a disaster, but it takes time to restore the system with all the correct settings in order to restore the functionality. This may not be easy if you have not recorded all the settings for the programs that use the data (see Applications Sharing Data below).

Backing up to a local system does not offer disaster protection. Do not backup to the same system, or at least make sure it is the only backup. Internet backup is available from many sources, but you need to check on the allowable size of the backups. Also, transmitting tens or hundreds of gigabytes could take days depending on your upload bandwidth. After the initial total backup, incremental backups will be faster. You should also determine if the internet server itself is backed up as well as make sure you know how to restore and the resources needed for the restore.

How Quickly Do You Need To Restore?
How quick depends on the functionality that has been lost. Having a corrupt file is completely different than having a corrupt data base. If the database is part of a POS in a high traffic area and it is December 22, then restoring will take too long. You may have to consider redundant systems as part of the Backup/Restore strategy. A small or lean operation may not have the surplus man hours to perform a restore. You may want to consider an MSP. But you should know up front the restore scenario and will it fit your business operation.

Applications Sharing Data
This may be your most difficult Backup/Restore scenario. The data in each application may need to be in synch to function properly or produce valid results. If that is the case, you need to restore using synced data sets or manually sync after restore. You can be in this situation if using programs such as an accounting program that obtains inventory data from a POS database. If the inventory data base must be restored and the backup is old, the accounting program may complain or produce inaccurate results. This situation used to be a problem for large enterprises but now small businesses may need to consider this issue. As we move data to the internet and share it with other applications, the timing and impact of a restore must also be considered.

Can Virtualization Help?
If your Backup/Restore or Disaster Recovery is not just about individual files and data is shared between applications, you may want to consider Virtualization. Virtualization itself is not a backup technology but it can be used to more easily to restore complex computer environments. The virtualized computer and its programs and data are a software image as a guest on a hosting computer. The entire image is backed up and restored as one file so it solves the problems of syncing data, restoring program configurations, and disaster recovery. The downside is that the image will be tens or hundreds of gigabytes big.

Proving The Restore
You need to go through the exercise of proving you can restore from your backups, something that should be done at least annually. This is not easy or convenient, but it is necessary and the Backup/Restore strategy should consider how to prove it works. Think of it as making sure the inflatable life raft really does inflate before you really do need it. Also, we recommend that you do not do a test restore to a production system.

Rich Rutkowski is owner/manager of Sale Point Data, LLC which has developed Distributed Sale Point, an internet-based full featured Point of Sale System. It is designed for businesses that must manage large or expensive inventory and provides encryption, high availability, and multi-lane and multi-store capability.

Mr. Rutkowski has more than 20 years of experience developing complex systems with large software content where safety of flight is a major consideration. His experience ranges from software developer, system tester, Engineering Manager, and Director of Engineering.