Just-in-Time Software Image Assembly for Data Center Storage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional data centers face storage management inefficiencies due to replicated software stacks across diverse machine configurations, leading to increased administrative and storage costs, especially exacerbated by virtual computing environments where many machines have different or identical operating systems, drivers, and applications.
Innovation Solution
Implementing 'just-in-time' image creation services that define required images based on workload policies, allowing for efficient assembly and disassembly of software stacks as workload-specific deltas, reducing the need for multiple copies of software stacks by storing only a single 'golden' image and assembling them on demand.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple copies of software stacks are stored for different machine configurations, then deployment flexibility is improved, but storage capacity is wasted and administrative complexity increases
Solution Approach 1:
The software stack is segmented into a common base image and workload-specific deltas. The base image contains shared components that can be reused across multiple configurations, while deltas contain only the unique portions needed for specific workloads. This segmentation allows the system to maintain deployment flexibility for different machine configurations while significantly reducing total storage requirements by avoiding redundant storage of identical software components.
2Adaptability or versatility
If multiple copies of software stacks are stored for different machine configurations, then deployment flexibility is improved, but administrative complexity increases
Solution Approach 1:
By segmenting software stacks into reusable base images and configurable deltas, the system simplifies administrative tasks. Patch management becomes easier since updates can be applied to the base image once and automatically propagated to all derived configurations. Image creation is simplified through automated assembly of base images and deltas, reducing the manual effort required to maintain multiple software stack versions.
Solution Approach 2:
The base image serves as a universal foundation that can be reused across multiple different workload configurations. This multi-functionality allows a single base image to support numerous derived images with different workloads, eliminating the need to maintain separate complete copies for each configuration and thereby reducing administrative overhead.
3Speed
If complete software stacks are pre-assembled and stored for each workload, then deployment speed is improved, but storage overhead increases
Solution Approach 1:
The base image is pre-assembled and stored once, containing all common software components that will be needed across multiple workloads. This preliminary preparation of the base image enables rapid deployment of derived images, as the system only needs to apply the much smaller workload-specific deltas rather than assembling complete software stacks from scratch each time.
Solution Approach 2:
The system transitions from storing complete software stacks in a two-dimensional space (each stack fully assembled) to a three-dimensional structure with a shared base layer and configurable delta layers. This dimensional change allows the base image to serve multiple derived images simultaneously, reducing storage overhead while maintaining the ability to quickly assemble complete images when needed.
Data Source
AI summary
Methods and apparatus involve “just-in-time” image creation services to address storage management issues faced by data centers. Representatively, a computing policy defines what images are required in software stacks according to intended workloads of computing devices during use and an order of assembly for each software stack is established. Upon a trigger event to instantiate an operating workload of any computing device, a workload image is configured for deployment to the hardware platform of the computing device, including assembling the stack in the order of assembly from stored virgin images of an operating system and any relevant applications. A delta of workload information is calculated when an operating workload for the computing device is to be brought down. Upon a trigger event to later re-instantiate the operating workload, another workload image is configured for deployment including combining the delta with the software stack, again, configured in the order of assembly.