All our servers are built with fully redundant hardware such as redundant storage arrays, redundant internal power supplies, error correcting memory, and a uninterruptible power source (battery backups). Redundant hardware allows for hardware errors or complete failure without interrupting the operations of the server.
In the event of a physical component failure we are able to replace the faulty components without having to power down the server, this allows your server to continue running without any interruption to your website, customers, or cloud based applications. Redundant hardware is key to dealing with physical failures and keeping our servers online 99.999% of the time.
Because multiple data storage, backup, and redundancy problems cannot be solved with hardware we had to utilise multiple software technologies to create the software redundancy component of our complete backup & disaster recovery strategy. Our software systems provide solution to multiple problems such as critical hardware failures, malicious software attacks, network errors, high traffic peak times, human error, deleted files recovery, accidental file modification, file transfer errors, and file corruption. Some of the technologies are included in real-time synchronization, shadow copy, duplicate server, and periodic backups.
Shadow copy provides you the ability to restore files back to an earlier point in time to reverse unwanted changes or human error. Every time a change is made to a file shadow copy takes a copy of the file, these copies are done completely behind the scenes without any user input.
A similar system is also applies to your server files and additionally your cloud based office programs.
We utilise a synchronization system due to some problem legacy backup technologies were causing. One of the major problems with FTP backups is they introduce a very sudden and large amount of network traffic every time the backup runs, if your server’s files are 119 gigabytes total and you run backups every 6 hours then this will cause the server to have 4 massive data spikes, these spikes may cause slow page loading times or even disconnect some services.
Slowing transfer speeds and reducing the frequency of backups are not a solution so we utilise a synchronization system that allows us to synchronize files between backup periods thus eliminating the sudden data spikes. Our backup server pauses synchronization during backups and resumes synchronization after the backup completes. Because communication between servers ceases during backups this actually results in network rate dip opposed to spikes seen with traditional FTP backups.
The chart is a log of network data transfer measured in gigabytes over a 24 hour period. The chart illustrates the difference between FTP and BTS backup methods.
- FTP Backup causes large network spikes and potentially results in slower loading times.
- BTS Backup method does causes overall network load to increase slightly but does not result in traffic spikes or any performance loss.
- BTS allows for the same amount of data to transfer over a longer time period opposes to FTP which has to transfer all data as quickly as possible.
The synchronization software splits files into multiple segments and calculates a checksum for each segment. This checksum system is used for multiple purposes, first it allows the software to detect which segments of a file have been modified and requires only those segments to be synchronized opposed to the entire file. Second the checksum allows for error detection, if a file is synchronized with recipient and results in a different checksum this means the file has been tampered with or corrupted during transmission; because the recipient is aware the file has transferred with errors it is able to correct the error immediately by requesting the file again. Almost all other file transfer or copying software will not detect corruption.
Periodic backups are still an essential component to a complete backup and disaster recovery solution even though you may be using redundant hardware. hardware redundancy does not provide any protection against critical failures, corrupt files, accidental deletion, human error, or malicious ransomware.
With our periodic backup service we will take frequent copies of all your data and keep each backup for 48 hours before it's deleted and replaced with a newer copy. The frequency of backups is upto you but generally 6 hourly, 12 hourly, or daily; and backups are usually kept for 48 hours however we can keep backups as long as you'd like.
Long term backups are useful for copy protection & records, legal, file revision longer than shadow copy, and etcetera.
A backup is taken every week, fortnight, month, 3 months, 6 months, and 1 year, however the only backups that are kept indefinitely is each yearly. Even though 24 weekly backups will be taken in a 6 month period only one will actually be kept for 6 months.
Backup and Disaster Recovery Checklist
- Power outages
- Physical drive failures
- Critical failures
- Network faults
- Malicious software
- Natural disasters
- Accidental changes
- Accidental deletion
- File corruption