Dynamic TCP Reassembly Context Allocation for Virtual Network Interfaces

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing virtual network interfaces struggle with dynamic TCP reassembly in virtualized environments, as they rely on statically assigned global reassembly contexts, which are not well-suited for the creation and destruction of virtual network interfaces, leading to inefficiencies and potential performance issues.

Innovation Solution

A method for dynamically allocating TCP reassembly contexts through a global common pool shared across all virtual network interfaces (VNICs), managed by a hypervisor that re-evaluates resource needs and allocates resources based on bandwidth ratios, preventing resource overuse by overly active VNICs and optimizing hardware resource usage.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If statically assigned global reassembly contexts are used, then hardware resources are simplified, but adaptability to virtualized environments deteriorates

Engineering Contradiction:

Improvehardware resourcesVSAvoidadaptability to virtualized environments

Core Design Contradiction:

Device complexityVSAdaptability or versatility

Solution Approach 1:

The patent implements dynamic allocation of TCP reassembly contexts where the number of contexts assigned to each VNIC can change based on current workload demands. The hypervisor continuously monitors packet reassembly needs and adjusts context allocation accordingly, allowing the system to adapt to varying virtualized environment requirements while maintaining hardware efficiency.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent creates a pool of shared TCP reassembly contexts that can be dynamically allocated to any VNIC that needs them. This universal pool serves multiple VNICs simultaneously, allowing the same hardware resources to be reused across different virtual network interfaces based on demand, rather than dedicating fixed contexts to each VNIC.

Inventive Principle:

Principle #6Universality (Multi-functionality)

2Adaptability or versatility

If dynamically created VNICs are supported, then adaptability improves, but resource management complexity increases

Engineering Contradiction:

Improvesupport for dynamically created VNICsVSAvoidresource management complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent implements a self-service mechanism where VNICs automatically request and receive TCP reassembly contexts from the shared pool when needed, and return contexts when no longer needed. The hypervisor's allocation logic automatically responds to these requests based on current pool availability and VNIC priorities, eliminating the need for manual resource management while supporting dynamic VNIC creation and destruction.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent incorporates feedback mechanisms where the hypervisor monitors the usage patterns of TCP reassembly contexts across VNICs and adjusts allocation dynamically. When a VNIC shows high packet reassembly activity, the system allocates more contexts to it; when activity decreases, contexts are released back to the pool. This feedback-driven approach simplifies resource management by automatically adapting to actual workload conditions.

Inventive Principle:

Principle #23Feedback

3Productivity

If more reassembly contexts are allocated, then TCP packet reassembly performance improves, but hardware resource usage increases

Engineering Contradiction:

ImproveTCP packet reassembly performanceVSAvoidhardware resources

Core Design Contradiction:

ProductivityVSQuantity of substance

Solution Approach 1:

The patent changes the parameter of context allocation from fixed to variable, allowing the number of active TCP reassembly contexts to fluctuate based on actual workload demands. During high-traffic periods, more contexts are allocated to maintain performance; during low-traffic periods, contexts are released to conserve hardware resources. This dynamic parameter adjustment optimizes the balance between performance and resource usage.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent merges the TCP reassembly context pools of multiple VNICs into a single shared pool managed by the hypervisor. Instead of each VNIC maintaining separate context allocations that may sit idle, the merged pool allows contexts to be freely shared and reused across all VNICs, maximizing hardware resource utilization while maintaining the ability to serve multiple virtual network interfaces simultaneously.

Inventive Principle:

Principle #5Merging (Combining)

Data Source

PatentUS9363193B2Virtualized network interface for TCP reassembly buffer allocation

Publication Date: 2016.06.07 MARVELL ASIA PTE LTD

AI summary

A method and a system embodying the method for dynamically allocating context for Transmission Control Protocol (TCP) reassembly, comprising providing a fixed plurality of global common TCP contexts; reserving for each of one or more virtual network interface card(s) one or more TCP context(s) out of the fixed plurality of the global common TCP contexts; and allocating to a virtual network interface card from the one or more virtual network interface card(s) a TCP context from the reserved one or more TCP contexts when a reassemble able TCP packet is received by the virtual network interface card, is disclosed.