Management Processor Peer-to-Peer Data Transfer via Communication Fabric
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
As networked storage systems and remote computing systems increase in density, physical limitations such as space requirements for network interconnect and environmental climate control become constraints, and traditional bulk storage systems are limited in the number of devices per host, hindering capacity, redundancy, and reliability in storage environments.
Innovation Solution
The implementation of a management processor that initiates communication arrangements between endpoint devices coupled to a communication fabric, allowing for peer-to-peer data transfers without passing through a host processor, thereby redirecting data transfers based on address ranges and utilizing PCIe fabric to logically isolate and route traffic among components.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If traditional bulk storage systems are used with host processors, then data transfers can be managed centrally, but the number of devices per host is limited and host processor resources are consumed
Solution Approach 1:
The system segments the storage environment into multiple compute units, each with its own host processor managing a limited set of endpoint devices. This segmentation allows the overall system to support many more devices than a single host could manage, while keeping each host's burden manageable. The communication fabric connects these compute units, enabling peer-to-peer transfers between endpoints in different compute units without involving the host processors.
Solution Approach 2:
The communication fabric acts as an intermediary between endpoint devices in different compute units. Instead of having host processors directly manage all device communications, the fabric enables direct peer-to-peer transfers by intercepting and redirecting transfers based on address ranges, eliminating the need for host processor involvement in these communications.
2Quantity of substance
If networked storage systems increase in density, then storage capacity increases, but physical space requirements for network interconnect and climate control become constraints
Solution Approach 1:
The patent merges storage and computing resources into integrated compute units, where endpoint devices are coupled to the communication fabric through host processors. This consolidation allows for higher density installations by combining previously separate infrastructure components into unified units that can be rack-mounted, reducing the overall physical footprint while maintaining high storage capacity.
3Productivity
If peer-to-peer transfers are enabled without host processor involvement, then transfer efficiency improves, but address routing complexity increases
Solution Approach 1:
The communication fabric provides self-service address routing capabilities through built-in address traps and redirection mechanisms. When an endpoint device initiates a transfer, the fabric automatically intercepts the transfer based on destination address ranges and redirects it to the appropriate endpoint without requiring host processor intervention. This self-service approach maintains high transfer efficiency while managing routing complexity within the fabric itself.
Data Source
AI summary
Computing architectures, platforms, and systems are provided herein. In one example, system is provided. The system includes a management processor configured to initiate a communication arrangement between a first endpoint device coupled to a communication fabric and a second endpoint device coupled to the communication fabric. The communication arrangement is configured to redirect a transfer from the first endpoint device based on an address corresponding to an address range of the second endpoint device without passing the transfer through a host processor coupled to the communication fabric that executes an application initiating the transfer.