Object Memory Fabric Across Shared Links
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current commodity hardware and software solutions are inadequate for meeting the demands of Cloud Computing and Big Data environments due to their complexity, inefficiency in managing processing, memory, and storage, leading to limitations in performance and scalability.
Innovation Solution
Implementing an object-based memory fabric that distributes and tracks memory objects across a hierarchy of processing nodes, using a highly threaded and latency-tolerant protocol, and eliminating distinctions between processing, memory, and storage by managing objects natively within the memory layer.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If processing, memory, and storage are managed separately by software in current commodity hardware, then system flexibility and ease of operation are maintained, but system complexity increases and performance efficiency deteriorates
Solution Approach 1:
The patent merges processing, memory, and storage management into a unified hardware-level system. The memory fabric integrates memory controllers, routing switches, and storage devices into a single coherent hierarchy, eliminating the need for separate software management layers while reducing overall system complexity through hardware-level integration.
Solution Approach 2:
The patent introduces a memory fabric intermediary layer between traditional processing and storage components. This memory fabric acts as a mediator that provides unified access to memory and storage resources, reducing the complexity of coordinating separate software-managed systems while maintaining operational flexibility.
2Adaptability or versatility
If separate processing, memory, and storage are managed by software on each node, then node independence is maintained, but bandwidth consumption and overhead increase
Solution Approach 1:
The patent segments the memory hierarchy into multiple levels (L1 cache, L2 cache, L3 cache, main memory, storage) with specialized controllers for each segment. This segmentation allows data to be cached and processed at different levels, reducing the need for high-bandwidth communication between all nodes while maintaining node independence through hierarchical access rights.
Solution Approach 2:
The patent implements a nested memory hierarchy where smaller, faster memory segments are contained within larger, slower ones. L1 cache is nested within L2, which is nested within L3, and so on through to storage. This nesting allows efficient data access at multiple granularities, reducing overall bandwidth consumption while preserving node independence through the nested structure.
3Productivity
If memory objects are distributed across a hierarchy of processing nodes, then performance and efficiency are improved, but tracking and managing memory objects becomes more complex
Solution Approach 1:
The patent implements feedback mechanisms at each level of the memory hierarchy where memory controllers continuously report the status and location of memory objects to higher-level controllers. This feedback enables automatic tracking and management of distributed memory objects without requiring complex centralized control, as each level manages its own objects while receiving feedback from lower levels.
Solution Approach 2:
The patent enables self-service management where each memory controller automatically tracks and manages its own memory objects without requiring external intervention. The hierarchical structure allows each level to independently manage its memory objects while automatically coordinating with other levels, reducing the complexity of centralized tracking while maintaining productivity improvements.
Data Source
AI summary
According to one embodiment, a hardware-based processing node of a plurality of hardware-based processing nodes in an object memory fabric can comprise a memory module storing and managing a plurality of memory objects in a hierarchy of the object memory fabric. Each memory object can be created natively within the memory module, accessed using a single memory reference instruction without Input/Output (I/O) instructions, and managed by the memory module at a single memory layer. The object memory fabric can utilize a memory fabric protocol between the hardware-based processing node and one or more other nodes of the plurality of hardware-based processing nodes to distribute and track the memory objects across the object memory fabric. The memory fabric protocol can be utilized across a dedicated link or across a shared link between the hardware-based processing node and one or more other nodes of the plurality of hardware-based processing nodes.