Template Virtual Machines for Memory Sharing

The patent applies preliminary action by pre-configuring template virtual machines with uniform code pages before deployment. The template VMs are prepared in advance with standardized operating systems and applications, allowing their code pages to be identified and shared across multiple instances without requiring dynamic monitoring during runtime. This advance preparation eliminates the need for continuous system resource consumption in monitoring page sharing operations.

The patent uses copying by creating multiple virtual machine instances from a single template VM. The template VM's code pages are copied and shared across multiple derived VMs, eliminating the need for each VM to maintain separate copies of identical code. This copying approach reduces physical memory requirements while minimizing overhead, as the sharing relationship is established through the template structure rather than dynamic monitoring.

The patent applies universality by creating a template VM that serves multiple functions across different virtual machine instances. The template VM contains standardized code pages that are shared universally across multiple derived VMs, allowing the same physical memory pages to serve multiple computational tasks. This multi-functionality approach enables higher consolidation ratios without proportionally increasing physical memory requirements.

The patent merges identical code pages from multiple virtual machine instances into shared memory spaces. By combining the code sections of multiple VMs that execute the same operating system and applications, the system reduces total memory consumption while maintaining the ability to run multiple VMs simultaneously, thus resolving the memory bottleneck that limits consolidation ratios.

The patent segments the virtual machine creation process into distinct phases: first creating the template VM without machine-specific identifiers to enable immediate memory sharing, then subsequently adding customizations to derived instances. This segmentation allows memory sharing to begin immediately while deferring customization to a later stage, resolving the contradiction between sharing efficiency and customization timing.

The patent applies preliminary action by preparing the base VM structure with shared code pages before adding machine-specific identifiers. The template VM is configured in advance with universal code that can be immediately shared, and customization is performed as a subsequent step when derived instances are created. This approach ensures memory sharing efficiency is achieved first, with customization following without blocking the sharing benefit.