5th December 2019: RSM Partners Senior Security Consultant Lennie Dymoke-Bradshaw's technical briefing paper examines the reasons and methods for performing backups of data stored on z/OS systems, and how that reasoning and thinking may change in light of the introduction of z/OS data set encryption. Data backups protect against several scenarios. We need to distinguish those situations from each other to ensure we protect against every eventuality.
Data can become corrupted, for example. This can result from programming errors, hardware failures, malicious intent, or process errors. For the backup to be useful, it may need to be physically separated from the master copy by some mechanism.
The data set containing the data can also be deleted: accidentally or, again, through malicious intent. The results are the same. As far as recovery is concerned, similar rules also apply. Separation of the backup copy from the master is important. Entire data centers can become non-operational or, in worst-case scenarios, be destroyed. In this situation, the physical separation of the backup copy from the master is essential. It's also especially useful if the copy is as current as possible, so synchronous copy mechanisms such as peer to peer remote copy PPRC can be used to great effect. At the same time, and this is not really a requirement for backup, archiving requirements should be considered separately. It's inevitable that some archival requirements can be met by examination of historical backup copies of data.
The important thing about backup copies of data is that data availability should not be subject to a single-event loss of the data.
Backup methods and media
Backup copies of data are performed using multiple mechanisms. These include access method copies, block copy mechanisms, real-time replication, flash copies, and DFSMShsm. And to be useful, backups should possess several qualities: availability, integrity, predictability and granularity. Anything that affects these qualities can affect the ability to recover data from a backup and hence affect the ability to recover the application requiring the data set. Types of media used vary from disk, tape and VTS virtual tape server devices. Each of these methods and media come with various risks attached.
Pervasive encryption technology
The recent technical briefing explains what can happen when you add z/OS data set encryption to the mix. This is part of the IBM technology called Pervasive Encryption that enables the use of encryption in multiple components of z/OS and Z Systems. Using z/OS data set encryption allows the data stored on the disk to not be readily understood by humans or an application program without first passing the data through a decryption process.
It's clear that encryption of z/OS data sets can provide a strong additional strategy to address certain security and compliance requirements. However, you also need to remember that encryption of data sets does not protect you against all threats. It does not protect you from data exfiltration by malicious actors who gain access to your keys, for example. Therefore, effective application key management and strict master key control are essential.
If you are implementing encryption of z/OS data sets, we are recommending a set of action points on how data should be replicated for backup purposes. These points-and many more-are covered in the technical briefing paper.
To download the technical briefing for free, please visit our literature Library here or for more information, email firstname.lastname@example.org