CPU Implemented Method for Backing Up Modified Data Sets in Non-Volatile Store for Recovery in the Event of CPU Failure

1. In a central processing unit (CPU) having a non-volatile store containing groups of logically independent non-fragmented, non-dispersed sequential datasets, and processor means, said processor means including means for updating and for backing up datasets in each group , said backing up occurring by copying said datasets into a first failure independent part of the nonvolatile store on a selective and on a periodic basis, said processor means further including timing means for assigning a maximum time interval (GBF) to each group in said store and for amending each dataset in each group with a date time stamp denoting dates and times of last backup and last update, a CPU implemented method for backing up of non-fragmented, non-dispersed sequential datasets in non-volatile store and for the recovery thereof in the event of CPU failure, comprising the steps of: (a) during a backup cycle, ascertaining those data sets in each group conforming to at least one of the following criteria: (1) the date of their last update lies between an instant date and the date of last backup AND the difference between the instant date and the date of last backup equals or exceeds a first predetermined minimum backup interval (MBF), or (2) the difference between the instant date and the date of last backup exceeds GBF, AND MBF equals or is less than GBF; (b) copying the ascertained data sets to said first failure independent part of said non-volatile store; and (c) during a recovery cycle, recopying into a second failure independent part of non-volatile store from the first part of non-volatile store said backed up datasets and groups.

2. A CPU implemented method for backup copying of non-fragmented , non-dispersed sequential datasets residing in non-volatile store, said CPU including date and timing means, said datasets forming logically independent storage groups, said method using an INCREMENTAL BACKUP policy where only new and changed data since the last backup will in turn be elsewhere copied in a first failure independent part of said non-volatile store, each data set having written therein its dates of last backup and update, comprising the steps of: (a) during a backup cycle, ascertaining those datasets within each group eligible for backup by evaluating the date of last change and the date of last backup, a dataset being included for backup copying if it conforms to at least one of the following limitations: (1) the date of the last update lies between an instant date and the date of last backup AND the difference between the instant date and the date of last backup equals or exceeds a first time interval (MBF), or (2) the difference between the instant date and the date of last backup exceeds a second time interval (GBF), AND MBF equals or is less than GBF; and (b) elsewhere copying the ascertained datasets to said first failure independent part of said non-volatile storage.

3. The method according to claim 2, wherein said INCREMENTAL policy being in contradistinction to either FULL VOLUME backup where all data whether or not changed since the last backup will in turn be elsewhere copied, or COMBINATION backup where FULL VOLUME backup is periodically invoked and INCREMENTAL backup invoked therebetween.

4. The method according to claim 2, wherein MBF is the minimum backup interval for said data sets, and GBF is the backup interval for a cognizant storage group.