The integrity question seems to be based on Data Domain internal integrity check.
We recently came across similar questions in a tender.
This was Symantec's response:
NetBackup Appliances leverage enterprise class hardware RAID controllers and all backup data is stored in a RAID 6 configuration with one or more hot spares. RAID 6 is a dual-parity architecture that offers a high-level of data redundancy and fault resiliency and is capable of protecting against two simultaneous disk failures without any manual intervention. All appliance system data is stored on mirrored RAID 1 configuration which offers 100 percent data redundancy.
To meet stringent data reliability and constant uptime requirements, NetBackup Appliances utilize Veritas File System, a technology that has for more than 25 years been the standard for Online Transaction Processing (OLTP) and critical application solutions. It has an inherent capability of data protection and file awareness and has been chosen by numerous organizations where the highest levels of data integrity and availability are required.
Veritas File System employs a circular intent log with all changes written to this intent log in a synchronous manner before they are periodically flushed out to their actual disk blocks. By using this intent log, Veritas File System can recover from system downtime in a fraction of the time. When restarted, Veritas File System simply scans the intent log, noting which file system changes had completed and which had not, and proceeds accordingly. Veritas File System was designed as a journaling file system and as such adds a great degree of both availability and integrity to the NetBackup Appliance as compared to devices built on top of older block based file systems that have later added or bolted on journaling technology.
Preventing data collisions
Data deduplication systems identify duplicate data by matching digital fingerprints against an existing database of protected data and while extremely rare, this fingerprinting methodology can generate a false match. The data deduplication technology found in the NetBackup Appliance eliminates this risk through a dual fingerprint approach, which adds another layer of protection on top of the base MD5 data identification. Potential data collisions are prevented by the combination of two hashes resulting in a new fingerprint that is neither MD5 nor SHA-x. The combined hash eliminates the chance of a collision and effectively protects the system against artificially generated MD5 collision files.
Critical element protection, NetBackup Appliances provide backup copies of critical elements by automatically maintaining dual copies of each critical deduplication element during key periods of transaction processing. If needed, these secure copies are used to facilitate a self-recovery inside the NetBackup Appliances, ensuring data consistency. During normal operations, the NetBackup Appliance ensures the consistency of the data written to and residing in the deduplication storage pool, thus guaranteeing all data referenced in the catalog can be successfully retrieved. Multiple verification and consistency checks exist to help prevent external causes, such as hardware and power failures, from leaving the appliance in an inconsistent state. These consistency checks can be configured to run at regular intervals to check for either the existence of file data, to validate the cyclic redundancy check (CRC) checksum for the data or to recalculate the individual deduplication fingerprints of each data segment. Should any potential data corruption be detected, self-recovery processes are run and alerts are sent to the administrator.
