Virtualized I/O Director for Server Resource Sharing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing server configurations are over-provisioned with resources like NICs and HBAs due to the need for peak capacity and multiple network connections, leading to waste, complexity, and inflexibility, as resources are typically dedicated to individual servers rather than being shared.
Innovation Solution
Implementing virtualized I/O resources, such as virtual NICs (vNICs) and virtual HBAs (vHBAs), which are shared across multiple servers via an Ethernet network, utilizing an I/O director to manage and provide quality-of-service guarantees, allowing dynamic reconfiguration without physical changes to servers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If resources are assigned on a per server basis to ensure adequate capacity, then reliability and performance are improved, but device complexity and cost increase due to over-provisioning
Solution Approach 1:
The patent implements virtualized I/O resources that can be dynamically allocated to multiple servers, allowing a single physical resource pool to serve multiple functions and servers. The I/O director virtualizes peripherals such as storage, network, and graphics resources, enabling these resources to be shared across different servers based on demand rather than being dedicated to individual servers.
Solution Approach 2:
The I/O director acts as an intermediary between servers and physical I/O resources. It virtualizes the peripherals and manages the allocation dynamically, serving as a mediator that decouples the direct connection between servers and physical resources. This intermediary layer enables flexible resource assignment while maintaining performance guarantees through virtualization technologies.
2Reliability
If redundant resources are configured for fault-tolerance, then reliability is improved, but loss of substance increases due to resource waste
Solution Approach 1:
The patent merges multiple I/O resources into a unified virtualized pool managed by the I/O director. Instead of having separate redundant resources for each server, the system combines resources into a shared pool that provides fault-tolerance through virtualization. The virtualized resources can be dynamically reallocated and replicated as needed, providing redundancy without duplicating physical hardware for each server.
Solution Approach 2:
The virtualized resource pool enables dynamic recovery and reallocation of resources. When failures occur, resources can be recovered and reassigned to different servers through the I/O director's management layer. This allows the system to maintain fault-tolerance while recovering from failures without permanent resource loss or waste.
3Reliability
If dedicated I/O resources are assigned to each server, then reliability is improved, but adaptability decreases due to inability to reconfigure
Solution Approach 1:
The patent implements dynamic resource allocation through virtualization. The I/O director can dynamically assign and reassign virtualized I/O resources to different servers based on changing requirements. This dynamic capability allows the system to adapt to different workloads and configurations while maintaining reliable resource availability through the virtualized pool.
Solution Approach 2:
The patent segments physical I/O resources into virtualized units that can be independently allocated and managed. The I/O director divides the physical resource pool into virtual resources that can be assigned to different servers as needed. This segmentation enables flexible reconfiguration while maintaining the underlying physical infrastructure for reliable operation.
4Reliability
If over-provisioned resources are allocated to handle worst-case scenarios, then reliability is improved, but productivity decreases due to low utilization
Solution Approach 1:
The virtualized resource pool enables I/O resources to serve multiple servers and functions simultaneously. Instead of having dedicated over-provisioned resources for each server, the same physical resources can be shared across multiple servers based on actual demand. This multi-functionality increases utilization while maintaining the capacity to handle peak loads through dynamic allocation.
Solution Approach 2:
The patent enables continuous utilization of I/O resources by dynamically allocating them to different servers based on workload demands. Rather than having resources idle during low-utilization periods, the virtualized pool continuously assigns resources to active workloads, ensuring that capacity is always productively used while maintaining the ability to scale to worst-case scenarios.
Data Source
AI summary
Provided is a novel approach for connecting servers to peripherals, such as NICs, HBAs, and SAS/SATA controllers. Also provided are methods of arranging peripherals within one or more I/O directors, which are connected to the servers over an Ethernet network. Such arrangement allows sharing the same resource among multiple servers.