Cloud Browser Isolation with GPU Forking for Low-Latency Rendering

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing cloud-based browser isolation techniques using pixel-based solutions suffer from a decline in user experience due to rendering on cloud servers, leading to latency and inefficiencies.

Innovation Solution

Implement cloud-based browser isolation using GPU forking, which involves initializing multiple virtual GPU instances within the cloud, distributing rendering tasks in parallel, and performing all read operations locally to minimize latency and enhance user experience.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If pixel-based cloud rendering is used for browser isolation, then security is improved by isolating the browser from the user device, but user experience deteriorates due to rendering latency on cloud servers

Engineering Contradiction:

ImprovesecurityVSAvoidrendering latency

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent creates virtual GPU instances that are software copies of physical GPU capabilities, allowing rendering operations to be performed in the cloud while maintaining compatibility with client-side GPU operations. This copying approach enables the client device to receive pre-rendered graphical content without experiencing the latency of real-time cloud rendering

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The system performs rendering operations in advance on the cloud server before transmitting the rendered content to the client device. By completing the rendering work beforehand, the system eliminates the need for real-time rendering during user interaction, thus reducing perceived latency while maintaining security isolation

Inventive Principle:

Principle #10Preliminary action

2Productivity

If rendering tasks are distributed across multiple virtual GPU instances, then rendering performance is improved through parallel processing, but system complexity increases due to workload distribution management

Engineering Contradiction:

Improverendering performanceVSAvoidsystem complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent divides rendering workloads into discrete tasks that can be independently distributed across multiple virtual GPU instances. Each virtual GPU instance handles specific rendering tasks in parallel, and the results are aggregated to form the final rendered output. This segmentation enables scalable performance improvement without proportionally increasing system complexity

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The virtual GPU instances are designed to be universal and interchangeable, with standardized interfaces and workloads that can be distributed across any number of instances. This multi-functionality allows the same rendering task to be executed on different virtual GPU instances without requiring complex instance-specific configurations, thereby reducing system complexity

Inventive Principle:

Principle #6Universality (Multi-functionality)

3Speed

If all read operations are performed locally within the cloud-based system, then network latency is reduced by eliminating round-trips, but data consistency challenges arise across distributed GPUs

Engineering Contradiction:

Improveoperation speedVSAvoiddata consistency

Core Design Contradiction:

SpeedVSStability of the object's composition

Solution Approach 1:

The patent introduces a central coordinator or intermediary component that manages data consistency across distributed virtual GPU instances. This intermediary handles synchronization protocols, coordinate read operations, and ensure that all GPUs access consistent data states without requiring constant network communication, thus maintaining both speed and consistency

Inventive Principle:

Principle #24Intermediary (Mediator)

4Ease of operation

If browser isolation is implemented through virtualized GPU, then user experience is improved by enabling local rendering, but resource allocation complexity increases for managing multiple virtual GPU instances

Engineering Contradiction:

Improveuser experienceVSAvoidresource allocation complexity

Core Design Contradiction:

Ease of operationVSDevice complexity

Solution Approach 1:

The patent implements dynamic resource allocation where virtual GPU instances can be created, scaled, and terminated based on real-time demand. The system automatically adjusts the number and capacity of virtual GPU instances to match user needs, eliminating the need for static, over-provisioned infrastructure and reducing long-term resource allocation complexity

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS20250217923A1Cloud-Based Browser Isolation with Graphics Processing Unit (GPU) Forking

Publication Date: 2025.07.03 ZSCALER INC

AI summary

Systems and methods for browser isolation with Graphics Processing Unit (GPU) forking includes initializing a plurality of virtual GPU instances within one or more servers of a cloud-based system; receiving a rendering request from a client device; analyzing the rendering request and determining a workload distribution across a plurality of virtual Graphics Processing Unit (GPU) instances based on the analyzing; executing rendering tasks across the plurality of virtual GPU instances and generating rendering instructions; and pushing the rendering instructions to the client device.