Compute-Centric Object Store Using Virtual Containers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud computing systems face inefficiencies due to the use of virtual machines (VMs) that rely on hardware emulation, leading to reduced performance and resource utilization, especially as the number of VMs increases and workloads vary, causing instability.
Innovation Solution
A compute-centric object storage system that utilizes a global operating system kernel to manage a pool of virtual guest operating systems, allowing for efficient assignment and operation of virtual operating system containers, which perform compute operations directly on objects within the object store without the need for complex hypervisor coordination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If virtual machines are used to provide multi-tenant computing systems, then system isolation and security are improved, but hardware emulation reduces performance and resource utilization efficiency
Solution Approach 1:
The patent uses virtualization to create virtual copies of computing resources (virtual machines, containers, sandboxes) that provide isolation while sharing physical hardware. Instead of dedicating separate physical systems to each tenant, the system creates virtual copies that can be dynamically allocated and managed, achieving both isolation and efficient resource utilization.
Solution Approach 2:
The patent implements a universal computing platform that can serve multiple tenants with different requirements through a single infrastructure. The system provides multi-functionality by supporting various virtualization approaches (VMs, containers, sandboxes) and workloads (batch processing, streaming, interactive) on the same physical hardware, eliminating the need for separate dedicated systems.
2Adaptability or versatility
If the number of virtual machines on a host machine increases, then multi-tenancy capability is improved, but coordination complexity increases and system stability decreases
Solution Approach 1:
The patent segments the virtualization management into hierarchical layers: a hypervisor layer handles low-level resource allocation and isolation, while upper-layer management systems handle workload coordination and tenant-specific configurations. This segmentation allows the system to scale to many tenants without proportionally increasing overall coordination complexity, as each layer manages only its specific concerns.
Solution Approach 2:
The patent introduces intermediary management layers and abstraction interfaces between the hypervisor and individual virtual machines. These intermediaries handle the coordination complexity by providing standardized interfaces for resource allocation, workload scheduling, and tenant management, reducing the direct coordination burden as the number of VMs increases.
3Reliability
If hardware emulation is used to provide virtual machines, then system isolation is improved, but resource utilization efficiency deteriorates
Solution Approach 1:
The patent employs virtualization to create lightweight virtual copies (containers, sandboxes) that provide isolation without full hardware emulation overhead. These virtual copies share the underlying hardware directly through controlled interfaces, maintaining security and isolation boundaries while eliminating the performance penalty of emulating entire hardware stacks for each tenant.
Data Source
AI summary
Systems and methods for providing a compute-centric object store. An exemplary method may include receiving a request to perform a compute operation on at least a portion of an object store from a first user, the request identifying parameters of the compute operation, assigning virtual operating system containers to the objects of the object store from a pool of virtual operating system containers. The virtual operating system containers may perform the compute operation on the objects according to the identified parameters of the request. The method may also include clearing the virtual operating system containers and returning the virtual operating system containers to the pool.