System Recovery—Breaking through the Dissimilar Hardware Restore Challenge
Hardware Failure is Inevitable To combat data erosion and system failures, backup procedures must be designed to address a hardware failure in a timely, cost-effective manner. When hardware must be replaced, a rapid recovery solution is critical. In the event of hardware failure, a bare-metal recovery can be automated or manual. Each approach has distinct advantages.
Automated system recovery Automated bare-metal recovery is designed for rapid, systematic recovery. With automation, procedures are more likely to be predictable and simple. The user will not require as much training, so this approach should also be more reliable than manual recovery. Automated recovery of a Windows® system does, however, have limitations. Because an operating system, with its unique configuration, is designed at the time of installation for a specific hardware device, a traditional automated recovery cannot take into account the dissimilar hardware components that are at the core of the new computer system.
The most problematic components are the Windows hardware abstraction layer (HAL), the kernel, and the mass storage controllers. When a Windows system boots, these three elements must be correctly assigned to the hardware otherwise Windows will not boot. It is less critical to resolve conflicts with other devices because, once loaded, Windows makes those devices easy to detect and install.
Manual system recovery Because of the limitations of automated recovery regarding restoration to dissimilar hardware, many users choose to reinstall the operating system manually. In fact, in the past, when a key hardware component (storage controller, motherboard, processor, or HBA) failed, manual recovery was the only viable approach. When the operating system is reinstalled manually, each of these items is detected and installed in a clean environment. The drawback is that the system must be configured entirely from scratch, wasting precious resources - and service packs and hot fixes must also be applied.
Before data restoration can begin in a manual system recovery, applications must be installed and configured and system settings set to match company standards. The complexity of this process is beyond ad hoc management techniques, and it requires strict controls and procedures.
When preparing for bare-metal recovery to dissimilar hardware, users commonly keep a journal to account for the changes that have occurred on the computer. This manual method of record-keeping not only is tedious, but it often fails to account for many system changes, as well. In addition, some administrators make the effort to capture a system's most recent "cold image" on the rare occasion when that system can be taken offline. These steps amount to a significant effort in planning; moreover, the recovery process is extremely slow (see figure 1).