Managed Backup ServicesShorter RTPOs by fast DISASTER recovery
Vorsorge ist besser als Nachsorge - Better safe than sorry
Always-On and high availability require intelligent backup processes allowing for a quick disaster recovery of systems and data. With our provided backup and disaster recovery technologies we can realize a Return to Production within 15 minutes for applications and data.
Our backup systems for virtual machine backups are powered by the Forward Incremental Forever technology, which is based on a one-time creation of a full backup, followed by incremental backups. This means, that only those data blocks that have been changed since the last backup routine are transmitted. Therefore, a daily incremental backup is much faster than a daily full backup. At defined recovery-points the full backup will automatically be merged with the incremental backups to create a new (synthetic) full backup.
Fail-safe and transaction consistent backups
In contrast to an inconsistent backup, data of a fail-safe backup will be saved at the same time, by which effectively a snapshot of the whole content of a filesystem is created. Because pending I/O-operations of open files will not be saved, special procedures are necessary during restoration. Especially with databases, consistency checks are required, which lead to longer recovery times. Single transactions might even become lost.
The backup software we are using allows for an application-consistent backup based on Microsoft VSS to create backups on image-level without the need for special agents to be used in the VM. That way, VSS-supported applications (such as Microsoft Active Directory, SQL and Exchange) can be backed up without interruptions. While data is copied, database transactions can continue as usual.
For applications that are not supported by VSS or running on Linux, we can implement so-called Pre-Freeze or Post-Thaw processes, which will start before or after a VM snapshot.
In order to consistently backup MySQL databases without interruptions, we have designed special agents based on the LVM snapshot technology. We are using a special software to back up Percona / XtraDB, which, combined with redo and transaction logs, is able to do a so-called Point-In-Time Recovery.
- Backups via dedicated Backup-LAN to storage systems located in a separate datacenter
- Daily system-backup based on Changed Block Tracking and Forward Incremental Forever technologies
- Application-consistent backup via Microsoft VSS, LVM snapshot and especially developed software
- Quick restoration of either complete VMs or single files and application objects
- Recovery of SQL databases on transaction leve
- Automatic check of the recoverability of every created backup
- Extensive monitoring of backups regarding execution and integrity
Instant recovery on image- and object-level
With Instant VM Recovery we are able to start a VM image directly at the requested restore point and migrate it into the productive cloud environment. During migration process the backup image remains write protected. Any changes made will be saved separately and merged afterwards when the VM has been copied completely to the operational storage.
Instead of a full recovery it is also possible to extract single objects such as files, mailboxes or data tables from the image and integrate them into the operating system.
By this method, the duration of a recovery can be significantly reduced compared to a classic recovery, which might as well last up to several hours.