Secure Migratable Architecture for Mainframe Workload Migration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing computing systems face challenges in securely and efficiently migrating mainframe workloads across different computing architectures due to architectural mismatches, leading to performance inefficiencies and limited flexibility, especially when moving virtualized systems from one platform to another.
Innovation Solution
A secure migratable architecture that utilizes a programmable circuit and memory with area descriptors to manage memory allocation and data storage across different computing architectures, allowing for seamless migration and execution of workloads by associating tags with memory areas, reducing the need for separate tag checks and enhancing security and performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If virtualization software is used to host mainframe workloads on commodity systems, then flexibility and adaptability are improved, but performance efficiency deteriorates due to architectural mismatches and translation overhead
Solution Approach 1:
The patent introduces an intermediary translation layer that mediates between mainframe instructions and commodity processor operations. This translator converts mainframe instruction sets into equivalent commodity processor instructions, enabling compatibility while maintaining performance through optimized translation routines and caching mechanisms that reduce the overhead of architectural mismatch.
2Adaptability or versatility
If monolithic virtualization software is used to migrate systems across platforms, then adaptability is improved, but migration complexity and service interruption increase
Solution Approach 1:
The patent segments the monolithic virtualization software into modular components that can be independently managed and migrated. This includes separating the translator module, virtual machine manager, and hosted workloads into discrete units that can be moved across platforms incrementally, reducing migration complexity and allowing for easier troubleshooting and maintenance.
3Adaptability or versatility
If virtualization software performs translation for each operation, then adaptability across different architectures is improved, but computational overhead increases
Solution Approach 1:
The patent implements preliminary action by pre-translating and caching frequently used mainframe instructions into commodity processor instructions before execution. The system maintains translation lookup tables and instruction caches that store pre-computed translation mappings, allowing the virtualized system to retrieve and execute instructions with minimal real-time translation overhead, thus reducing computational energy consumption while maintaining architecture compatibility.
Data Source
AI summary
Methods and systems for implementing a secure migratable architecture having improved performance features over existing virtualization systems are disclosed. One method includes allocating a portion of a memory for use by a process, the process including a firmware environment representing a virtual computing system having a second computing architecture different from a first computing architecture of a computing system on which the process is executed. The method includes associating area descriptors with each of a plurality of memory areas within the portion of the memory used by the process, and receiving a request within the firmware environment to store data within a first memory area of the plurality of memory areas, the first memory area defined by a first area descriptor of the area descriptors, the request being associated with a plurality of memory addresses within the first memory area. The method includes, in response to the request, performing a check on a tag associated with the first memory area and stored in the first area descriptor. The method further includes, upon completion of the check, storing the data within the first memory area without performing a separate tag check for each of the plurality of memory addresses within the first memory area.