At this insurance company, disaster recovery policies and procedures are in place, both on the data center level and for the departments. Also, backup systems are also in place, again, on the department and data center levels. Depending on the value that specific data has to the business, disk-mirroring is used to protect some data.
In the real world, the 'pretty picture' ends here and things become messy. Even in the data center we often find more than just one open system backup/recovery solution. Typically, additional products can be also found for backup/recovery in the departmental server rooms. In some places, we even find more than one backup/recovery product in the same department, and in many places, the number of different backup/recovery products in use is near the number of departmental server rooms.
Now consider a disastrous event in such an environment. The data center will have a hot site, either internally or with an external service provider. Some of the departments will also have such a service set up for some of their servers.
But if, for example, the sprinkler goes off in the server room of the marketing department (the server room was set up for an interim period, but that was two years ago), systems and data will get lost. And if a nearby river floods more than the expected area (see Europe this summer of 2002), all systems and data are at risk.
If the data center is prepared for such an event, the switch-over to the disaster recovery site happens quickly and mainframe operations resume with very little delay. The situation is different for the open systems servers, though. Those highly critical systems, which were mirrored to the departmental hot site will go on-line fairly quickly, too. The others wont. Also, consider the number of people and the co-ordination effort required to achieve even this level of recovery.
On the mainframe side, a staff of two can handle disaster recovery for all applications and data. In order to speed up the recovery of the different open systems, the ratio of administrators to systems is much higher.
If the open systems data had been backed up to the mainframe, it would have been available at the data center disaster recovery site and servers could be rebuilt there, quickly, managed by a small staff, and according to set policies which adequately reflect the business relevance of specific data and applications.
Such a solution is coordinated and managed by default because all recovery is based on a single system.
The Meta Group describes the difference between data center and open systems backup like this: "....mainframe data centers have backup procedures down to a science, with elaborate policies, procedures, storage sites, [and] proven backup tools..." Whereas "..outside mainframe environments, backup processes remain spotty, with highly critical servers being the only universally guaranteed assets." (Source: Meta Group Market Study Disaster Recovery and Business Continuity Planning: Key to Corporate Survival)
So, when you're faced with consolidating backup and recovery for an enterprise, include the mainframe in the solution. Like that kid's big brother, the big iron will be there when you need it most.
About the Author : Christian Traue is the Director of Product Management for Tantia Technologies