KubeEdge

Term 3 · Sep-Nov · 4 projects

• Comprehensive Example Restoration for KubeEdge Ianvs: Phase IV

  LFX Mentorship · Listed in the official cohort

  @MooreZhengSource IssueLFX portal

  Ianvs serves as the KubeEdge SIG AI distributed benchmark toolkit. As more and more contributors running, KubeEdge Ianvs now has up to 30 examples, and the number is still increasing. KubeEdge Ianvs then faces mounting usability issues due to dependency evolution and validation mechanisms. As Python versions, third-party libraries, and Ianvs features advance, partial historical examples fail to execute. This has led to surging user-reported Issues from confused contributors, untested PRs breaking core functionality of legacy features, and severely outdated documentation misaligning with actual capabilities. Without systematic intervention, the example risks becoming obsolete for edge-AI developers and especially newcomers. We then try to resurrect Ianvs’ usability with a comprehensive example restoration.

  • Python
  • Benchmark
  • KubeEdge-Ianvs
  • AI/ML

• Enable RuntimeClass and Confidential Containers on KubeEdge

  LFX Mentorship · Listed in the official cohort

  @HongbingZhangSource IssueLFX portal

  KubeEdge does not yet provide complete RuntimeClass support for workloads running on edge nodes. This prevents users from selecting alternative runtime handlers such as Kata Containers and limits the use of confidential or hardware-isolated workloads at the edge. Kubernetes RuntimeClass provides a standard way for a Pod to select a runtime handler through spec.runtimeClassName. This project will implement and validate the missing end-to-end RuntimeClass workflow in KubeEdge, covering resource synchronization, edge-side caching, runtime handler resolution, error handling, restart recovery, and cloud-edge reconnection. The first stage will validate RuntimeClass with Kata Containers in a reproducible KubeEdge environment. After the basic workflow is completed, the project will integrate and validate a Confidential Containers runtime. Intel TDX or another supported confidential computing platform may be used when suitable infrastructure and community support are available. The project focuses on RuntimeClass and confidential workload enablement. It does not require developing a new container runtime, rebuilding the Confidential Containers stack, or implementing a complete attestation service.

  • Go
  • Kubernetes
  • KubeEdge
  • RuntimeClass
  • Containerd
  • Container Runtime Interface
  • +4

• KubeEdge-Ianvs Simulation Sandbox: Environment-Isolated Execution

  LFX Mentorship · Listed in the official cohort

  @MooreZhengSource IssueLFX portal

  For most distributed AI scheme developers, establishing and deploying a large-scale cloud-edge collaborative system is often complicated and burdensome. While KubeEdge-Ianvs currently provides a single-node algorithm tester to evaluate accuracy-wise metrics using test datasets, measuring system-wise metrics, such as bandwidth, computing power, and peak memory, in a real-world manner for large-scale nodes is extremely difficult and costly. Furthermore, executing all test cases within a single shared Python process easily triggers dependency conflicts, path contamination, and fatal OOM crashes when heavy-load models like LLMs, VLAs, and foundation models run alongside lightweight examples. To address these challenges, this project would like to introduce an industrial-grade distributed collaborative system simulation using a worker-in-worker approach on a single machine, providing low-cost, scalable testing capabilities, robust environment isolation, and precise system-level metric profiling.

  • KubeEdge-Ianvs
  • KubeEdge-Sedna
  • KubeEdge
  • Kubernetes
  • Docker
  • Linux Kernel mechanisms
  • +4

• Modernize KubeEdge Controllers and Admission Webhooks

  LFX Mentorship · Listed in the official cohort

  @DoisLONGSource IssueLFX portal

  KubeEdge currently uses separate implementation and deployment models for cloud-side controllers and admission webhooks. The controller manager already uses controller-runtime and reconciliation, while admission remains an independent command, HTTP server, Deployment, Service, and manually registered webhook configuration. Modern Kubebuilder and controller-runtime projects can run reconciliation controllers and admission webhooks through the same Manager, sharing lifecycle, client, scheme, cache, logging, health checks, leader election, metrics, certificates, and deployment management. This project will modernize the KubeEdge controller framework by upgrading the Go and Kubebuilder-related toolchain, aligning controller-runtime and controller-tools with KubeEdge's Kubernetes dependencies, migrating existing admission handlers to controller-runtime webhooks, and consolidating controller and admission capabilities into one controller manager component. It will also add automated vulnerability checks and improve toolchain version consistency.

  • Go
  • Kubernetes
  • KubeEdge
  • KubeBuilder
  • controller-runtime
  • controller-tools
  • +5

Term 2 · Jun-Aug · 5 projects

• Building an Edge-Cloud Collaborative Framework for Embodied AI Applications with KubeEdge

  LFX Mentorship · Listed in the official cohort

  @HongbingZhangSource IssueLFX portal

  This project aims to explore how KubeEdge can support real-world embodied AI applications, where intelligent devices such as robots, inspection systems, or autonomous mobile platforms need to perceive, process, and respond in edge environments. The mentee will select a representative embodied AI scenario and build an end-to-end edge-cloud collaboration solution with KubeEdge. The solution should cover device or data source access, edge-side data processing, lightweight AI inference, cloud-edge synchronization, and system behavior under unstable network conditions. Through this project, the community will gain a reproducible reference implementation that demonstrates how KubeEdge can be used as the infrastructure foundation for embodied AI workloads.

  • Kubernetes
  • KubeEdge
  • Go
  • Python
  • Edge Computing
  • Edge AI

• Comprehensive Example Restoration for KubeEdge Ianvs: Phase III

  LFX Mentorship · Listed in the official cohort

  @MooreZhengSource IssueLFX portal

  Ianvs serves as the KubeEdge SIG AI distributed benchmark toolkit. As more and more contributors running, KubeEdge Ianvs now has up to 30 examples, and the number is still increasing. KubeEdge Ianvs then faces mounting usability issues due to dependency evolution and validation mechanisms. As Python versions, third-party libraries, and Ianvs features advance, partial historical examples fail to execute. This has led to surging user-reported Issues from confused contributors, untested PRs breaking core functionality of legacy features, and severely outdated documentation misaligning with actual capabilities. Without systematic intervention, the example risks becoming obsolete for edge-AI developers and especially newcomers. We then try to resurrect Ianvs’ usability with a comprehensive example restoration.

  • Python
  • Benchmark
  • KubeEdge-Ianvs
  • AI/ML

• Enabling Edge-Native Inference for Lightweight Large Language Models with KubeEdge

  LFX Mentorship · Listed in the official cohort

  @DoisLONGSource IssueLFX portal

  This project focuses on validating KubeEdge as an edge-native infrastructure for lightweight large language model inference. As more AI services move from the cloud to edge devices, lightweight LLMs provide a practical way to reduce latency, protect data privacy, and support local intelligence under limited resources. The mentee will deploy one or more lightweight LLMs on KubeEdge edge nodes and evaluate the complete workflow, including model packaging, workload scheduling, service exposure, lifecycle management, and performance measurement. The project is expected to produce a practical reference for running generative AI workloads on edge nodes managed by KubeEdge.

  • Kubernetes
  • KubeEdge
  • Python
  • Go
  • LLM Inference
  • Edge AI
  • +1

• Exploring AI PCs as KubeEdge-Managed Edge Nodes for Intelligent Workload Orchestration

  LFX Mentorship · Listed in the official cohort

  @Shelley-BaoYueSource IssueLFX portal

  This project explores how AI PCs can be integrated into the KubeEdge ecosystem as a new type of edge node. With built-in heterogeneous computing capabilities such as CPUs, GPUs, and NPUs, AI PCs are becoming suitable platforms for local AI inference, privacy-sensitive applications, and low-latency intelligent services. The mentee will connect one or more AI PCs to KubeEdge, deploy representative AI workloads, and explore resource monitoring and workload orchestration patterns for AI PC environments. The project will summarize practical experience and provide reference documentation for using KubeEdge to manage AI PCs as edge computing infrastructure.

  • Kubernetes
  • KubeEdge
  • Edge AI
  • AI PC
  • Heterogeneous Computing
  • Resource Monitoring
  • +2

• Exploring Alternatives to iptableManager: Optimizing Edge-Cloud Request Forwarding via Apiserver Redirection to cloudcore

  LFX Mentorship · Listed in the official cohort

  @luomengYSource IssueLFX portal

  Currently, in KubeEdge, the cloudstream component utilizes iptables to redirect requests from the apiserver intended for the edge node's KubeletEndpoint to port 10003 of the cloudcore to which that edge node is connected. The request is then routed through cloudstream to edgestream, ultimately accessing the KubeletEndpoint port on the edged. This requires the iptableManager to use iptables for interception and forwarding every 10 seconds. Since this mechanism depends on iptables, a proposal has been made to explore whether it's possible to have the Apiserver redirect to cloudcore directly, thereby replacing the iptableManager.

  • Go
  • Kubernetes
  • Client-go

Term 1 · Mar-May · 4 projects

• Batch Enable/Disable of EdgeHub on Edge Nodes via kubeedge/keadm

  LFX Mentorship · Listed in the official cohort

  @luomengYSource IssueLFX portal

  In bandwidth limited environments, as the cluster size expands, communication between cloud and edge nodes in KubeEdge consumes a significant amount of network resources, which may interfere with upper layer business applications. To alleviate this issue, it is necessary to implement batch enable or disable the EdgeHub edge node feature in keadm. Administrators can temporarily take edge nodes offline by turning off the EdgeHub switch, thereby releasing the bandwidth required for critical services. This feature should support adding subcommands in keadm for implementation.

  • Go
  • KubeEdge

• Cloud-Edge Simulation Benchmark for LLM Speculative Decoding in KubeEdge-Ianvs

  LFX Mentorship · Listed in the official cohort

  @hsj576Source IssueLFX portal

  LLM inference acceleration is increasingly important for cloud-edge collaborative AI deployments. Speculative decoding can improve end-to-end generation speed by using a lightweight draft model to propose token candidates and a larger target model to verify them, but its real-world gains depend heavily on cloud-edge constraints such as network latency, bandwidth limits, and heterogeneous compute. Ianvs provides a unified benchmarking framework, and KubeEdge scenarios often require evaluating AI workloads under cloud-edge conditions. This project proposes a single-host cloud-edge simulation benchmark in Ianvs to evaluate speculative decoding for LLM inference. The benchmark will simulate edge (draft) and cloud (verify) roles as separate processes, inject configurable network constraints, and report standardized throughput and latency metrics, enabling reproducible comparison between baseline decoding and speculative decoding under different cloud-edge budgets.

  • Python
  • PyTorch
  • HuggingFace Transformers
  • LLM inference/decoding
  • benchmarking & performance profiling
  • KubeEdge/Ianvs basics

• Comprehensive Example Restoration for KubeEdge Ianvs

  LFX Mentorship · Listed in the official cohort

  @MooreZhengSource IssueLFX portal

  Ianvs serves as the KubeEdge SIG AI distributed benchmark toolkit. As more and more contributors running, KubeEdge Ianvs now has 25 examples, and the number is still increasing. KubeEdge Ianvs then faces mounting usability issues due to dependency evolution and validation mechanisms. As Python versions, third-party libraries, and Ianvs features advance, partial historical examples fail to execute. This has led to surging user-reported Issues from confused contributors, untested PRs breaking core functionality of legacy features, and severely outdated documentation misaligning with actual capabilities. Without systematic intervention, the example risks becoming obsolete for edge-AI developers and especially newcomers. We then try to resurrect Ianvs’ usability with a comprehensive example restoration.

  • Python
  • Benchmark
  • KubeEdge-Ianvs
  • AI/ML

• Enable and Verify KubeEdge Support on RISC-V Architecture

  LFX Mentorship · Listed in the official cohort

  @HongbingZhangSource IssueLFX portal

  With the rapid growth of the RISC-V ecosystem, an increasing number of eage computing devices and IoT boards are adopting RISC-V. In very early versions of KubeEgde, we briefly verified compatibility with RISC-V. However, this has not been continuously validated in recent releases. As the codebase has evolved significantly, we need to re-establish and solidify support for this architecture to ensure KubeEdge can run seamlessly on the next generation of open hardware.

  • KubeEdge
  • Operation System

Term 3 · Sep-Nov · 5 projects

• Comprehensive Example Restoration for KubeEdge Ianvs

  LFX Mentorship · Listed in the official cohort

  @MooreZhengSource IssueLFX portal

  Ianvs serves as KubeEdge SIG AI distributed benchmark toolkit. As more and more contributors running, KubeEdge Ianvs now has 25 examples and the number is still increasing. KubeEdge Ianvs then faces mounting usability issues due to dependency evolution and validation mechanisms. As Python versions, third-party libraries, and Ianvs features advance, partial historical examples fail to execute. This has led to surging user-reported Issues from confused contributors, untested PRs breaking core functionality of legacy features, severely outdated documentation misaligning with actual capabilities. Without systematic intervention, the example risks becoming obsolete for edge-AI developers and especially newcomers. We then try to resurrect Ianvs’ usability with comprehensive example restoration.

  • Python
  • Benchmark
  • KubeEdge-Ianvs
  • AI/ML

• Deep integration of KubeEdge with AMD edge nodes

  LFX Mentorship · Listed in the official cohort

  @HongbingZhangSource IssueLFX portal

  AMD chips, with their powerful x86 architecture, exceptional computing performance, and advanced NPUs, demonstrate significant potential in fields such as industrial automation, in-vehicle systems, and high-performance edge computing. Introducing AMD's robust general-purpose and heterogeneous computing capabilities into the KubeEdge ecosystem is crucial for handling increasingly complex and latency-sensitive edge AI applications.However, the deep integration, performance optimization, and best practices between KubeEdge and AMD's high-performance edge platforms—particularly their built-in NPUs and other hardware acceleration units—still require systematic exploration and validation. This project aims to establish a complete link between KubeEdge and AMD edge nodes, building a comprehensive edge computing solution from hardware deployment to NPU acceleration, thereby greatly enriching KubeEdge's hardware ecosystem.

  • Go
  • Kubernetes
  • KubeEdge
  • Linux
  • Hardware Integration
  • AI/ML

• Device Anomaly Detection Framework Based on KubeEdge

  LFX Mentorship · Listed in the official cohort

  @meixiezichuanSource IssueLFX portal

  The current KubeEdge platform represents device states using three statuses: Desired, ObservedDesired, and Reported. The device states displayed on the platform are entirely reliant on the Mapper, which collects and reports data from the device side. However, due to limitations in the Mapper implementation, physical device malfunctions, network delays, and potential network attacks, the device states shown on the platform may not accurately reflect the actual state of the devices. In the KubeEdge platform, if applications depend on device states for decision-making, such inconsistencies in state representation may lead to undesirable outcomes. Therefore, this project aims to design a device state anomaly detection framework for KubeEdge. By exploring the causal relationships among device states, the framework will establish lightweight anomaly detection capabilities and provide a comprehensive toolchain encompassing data collection, model training, real-time anomaly detection, and results visualization.

  • KubeEdge
  • IoT
  • Machine Learning

• Industrial Embodied Intelligence Benchmarking Dataset for KubeEdge-Ianvs

  LFX Mentorship · Listed in the official cohort

  @MooreZhengSource IssueLFX portal

  As industrial manufacturing accelerates its digital transformation through advancements in robotics, adaptive production lines, and smart testing systems, cloud-edge collaboration has emerged as a critical enabler for deploying embodied intelligence in complex operational environments. Contemporary industry requirements for embodied intelligence now extend beyond basic task execution to encompass multimodal perception and decision integration, dynamic environment adaptation, and distributed device orchestration. Existing benchmarking frameworks exhibit limitations in evaluating scenario-specific embodied attributes inherent to industrial settings. This initiative leverages the KubeEdge-Ianvs collaborative AI framework, integrating domain-specific test datasets, simulation environments, and quantitative metrics to establish a certified industrial-grade evaluation infrastructure for embodied intelligence systems.

  • Python
  • Benchmark
  • Dataset
  • Embodied Intelligence

• Research on Deploying Small Language Models with KubeEdge and Integrating with Enterprise AI Platforms

  LFX Mentorship · Listed in the official cohort

  @HongbingZhangSource IssueLFX portal

  KubeEdge, as a native edge computing platform built on the Kubernetes ecosystem, offers capabilities such as reliable cloud-edge communication, edge autonomy, and IoT device integration. However, its ability to support intelligent model execution at the edge has yet to be systematically validated and practiced in real-world scenarios. This research aims to explore the feasibility and performance of deploying and running small language models on edge nodes using KubeEdge

  • KubeEdge
  • LLM
  • Go
  • Python

Term 2 · Jun-Aug · 4 projects

• Automatically generate unit tests and e2e tests based on LLM

  LFX Mentorship · Listed in the official cohort

  @liyuerichSource IssueLFX portal

  Unit test and e2e test coverage is a critical metric for ensuring code quality and stability. However, many codebases suffer from low test coverage, which increases the risk of release quality, undetected bugs and regressions. We can leverage Large Language Models (LLMs) like DeepSeek to automatically generate unit tests/e2e tests for code with low coverage, then integrate this into CI/CD pipelines to submit Pull Requests (PRs) for people reviewing.

  • KubeEdge
  • LLM
  • Go
  • DevOps

• Device Anomaly Detection Framework Based on KubeEdge

  LFX Mentorship · Listed in the official cohort

  @meixiezichuanSource IssueLFX portal

  The current KubeEdge platform represents device states using three statuses: Desired, ObservedDesired, and Reported. The device states displayed on the platform are entirely reliant on the Mapper, which collects and reports data from the device side. However, due to limitations in the Mapper implementation, physical device malfunctions, network delays, and potential network attacks, the device states shown on the platform may not accurately reflect the actual state of the devices. In the KubeEdge platform, if applications depend on device states for decision-making, such inconsistencies in state representation may lead to undesirable outcomes. Therefore, this project aims to design a device state anomaly detection framework for KubeEdge. By exploring the causal relationships among device states, the framework will establish lightweight anomaly detection capabilities and provide a comprehensive toolchain encompassing data collection, model training, real-time anomaly detection, and results visualization.

  • KubeEdge
  • IoT
  • Machine Learning

• Embodied Intelligence Benchmarking Framework for Industrial Manufacturing with KubeEdge-Ianvs

  LFX Mentorship · Listed in the official cohort

  @MooreZhengSource IssueLFX portal

  As industrial manufacturing accelerates its digital transformation through advancements in robotics, adaptive production lines, and smart testing systems, cloud-edge collaboration has emerged as a critical enabler for deploying embodied intelligence in complex operational environments. Contemporary industry requirements for embodied intelligence now extend beyond basic task execution to encompass multimodal perception-decision integration, dynamic environment adaptation, and distributed device orchestration. Existing benchmarking frameworks exhibit limitations in evaluating scenario-specific embodied attributes inherent to industrial settings. This initiative leverages the KubeEdge-Ianvs collaborative AI framework, integrating domain-specific test datasets, simulation environments, and quantitative metrics to establish a certified industrial-grade evaluation infrastructure for embodied intelligence systems.

  • Python
  • Benchmark
  • Dataset
  • Embodied Intelligence

• Support KubeEdge EdgeNode Runing on RK3588 Chip

  LFX Mentorship · Listed in the official cohort

  @HongbingZhangSource IssueLFX portal

  The RK3588 chip, developed by Rockchip, is widely used in edge computing devices due to its balanced computational power, rich interface options, and low power consumption. Supporting RK3588 edge devices is crucial for expanding the KubeEdge ecosystem. However, it has not yet been fully validated whether RK3588-based edge nodes can be seamlessly integrated with KubeEdge. This project aims to establish complete compatibility between RK3588 and KubeEdge.

  • KubeEdge
  • Hardware
  • Go

Term 1 · Mar-May · 5 projects

• Community Website Comprehensive Upgrade Project: Homepage Renewal and Expansion of Core Pages

  LFX Mentorship · Listed in the official cohort

  @HongbingZhangSource IssueLFX portal

  To improve the user experience of the KubeEdge official website, this project will focus on homepage design enhancements, the addition of new pages, and improvements to community resources. The goal of this project is to enhance the website's usability, increase user engagement, and attract more users to KubeEdge by enhancing training content and hardware compatibility support.

  • KubeEdge
  • JavaScript
  • Docusaurus

• Domain-specific large model benchmarks: the edge perspective

  LFX Mentorship · Listed in the official cohort

  @MooreZhengSource IssueLFX portal

  Based on existing datasets, the issue aims to build an advanced benchmark for edge-oriented domain-specific large models on KubeEdge-Ianvs. It aims to help all Edge AI application developers validate and select the best-matched domain-specific large models. For Edge AI service providers, it also helps identify which scenarios, edge nodes, or even locations could have the best performance or improvement for their models.

  • KubeEdge-Ianvs
  • Python
  • LLMs

• Enhance Dependency Management and Documentation for KubeEdge-Ianvs

  LFX Mentorship · Listed in the official cohort

  @FuryMartinSource IssueLFX portal

  Ianvs is currently grappling with significant dependency management challenges. It lacks a robust system to handle updates and ensure compatibility. As Python versions, dependency libraries, and Ianvs features continuously evolve, many existing examples fail to run, resulting in a surge of inquiries in the Issues section. Moreover, new PRs are often merged without being tested against historical examples, making it difficult to guarantee the functionality of past features through manual Code Review alone. There is an urgent need for a more comprehensive CI testing framework to maintain the usability of Ianvs features as the project progresses. Additionally, the online documentation is outdated, which can be quite confusing for new users.

  • KubeEdge
  • Ianvs
  • Python
  • CI/CD pipelines

• Enhance KubeEdge testing coverage

  LFX Mentorship · Listed in the official cohort

  @wbc6080Source IssueLFX portal

  KubeEdge would like to improve the UT coverage of the code to better maintain the quality of the code and reduce the introduction of defects. Increase the UT coverage rate to 60% to 70% (currently, the UT coverage rate is 38.69% ). It is important to note that in addition to requiring the overall UT coverage of KubeEdge to meet the requirements, the UT coverage of each core code directory(cloud/, edge/, keadm/ and pkg/) also needs to exceed 60%.

  • KubeEdge
  • Go
  • Testing

• KubeEdge Dashboard Enhancement - BFF

  LFX Mentorship · Listed in the official cohort

  @ghosindSource IssueLFX portal

  To improve the performance of KubeEdge dashboard, we would like to introduce a BFF (Backend for Frontend) layer. It serves as a middle layer to handle the communication between the dashboard and the KubeEdge API, providing a more efficient, secure, and maintainable solution.

  • KubeEdge
  • JavaScript
  • React

Term 3 · Sep-Nov · 5 projects

• Cloud-edge collaborative speculative decoding for LLM based on KubeEdge-Ianvs

  LFX Mentorship · Listed in the official cohort

  @hsj576Source IssueLFX portal

  The autoregressive decoding mode of LLM determines that LLM can only be decoded serially, which limits its inference speed. Speculative decoding technique can be used to decode LLM in parallel with the help of draft model, so as to improve the inference speed of LLM without loss of accuracy. However, the speculative decoding technology of LLM does not consider the application in the cloud-edge distributed environment. This project aims to implement cloud-edge collaborative speculative decoding based on KubeEdge-Ianvs, an open source cloud-edge collaborative distributed machine learning platform, so as to further improve the LLM inference speed in cloud-edge environment.

  • KubeEdge-Ianvs
  • LLM
  • Pytorch
  • Python

• Decouple the node cooperation ability and batch management ability of the edgeapplication

  LFX Mentorship · Listed in the official cohort

  @WillardHuSource IssueLFX portal

  EdgeApplication can be overrides deployment spec(i.e. replicas, image, commands and environments) via the node group, and pod traffics are closed-loop in a node group(Deployments managed by EdgeApplication share a Service). But in the real scenario, the scope of nodes that need batch operations is different from that of nodes that need to collaborate with each other. Therefore, we need to have a solution to decouple the node cooperation ability and batch management ability of the edgeapplication.

  • Kubernetes
  • KubeEdge
  • Go

• Elastic Inference for Deep Learning Models Using KubeEdge

  LFX Mentorship · Listed in the official cohort

  @tangming1996Source IssueLFX portal

  The rapid advancement of AI has led to the widespread application of deep learning models across various fields. However, the resource demands for model inference tasks can fluctuate significantly, especially during peak periods, posing a challenge to the system's computing capabilities. To address this varying load demand, we propose an elastic inference solution leveraging KubeEdge and Horizontal Pod Autoscaling (HPA) to enable dynamic scaling of inference tasks. By utilizing KubeEdge, we can distribute inference tasks across different edge devices and cloud resources, achieving efficient resource utilization and task processing.

  • KubeEdge and its subproject Sedna frameworks.
  • Experience in deploying and managing Kubernetes
  • including configuring and tuning the HPA mechanism.
  • Expertise in developing and tuning deep learning models.
  • Programming experience
  • particularly in Python and Go.

• Integrate KubeEdge, Sedna, and Volcano for High-Performance Training Task Scheduling

  LFX Mentorship · Listed in the official cohort

  @Shelley-BaoYueSource IssueLFX portal

  KubeEdge and Sedna have already enabled edge-cloud collaborative training and collaborative inference capabilities. We aim to explore and foster collaborations with more communities to provide enhanced AI capabilities. By integrating Volcano, we aim to achieve high-performance scheduling within the cloud-edge collaborative framework, thereby pushing the boundaries of what can be achieved in distributed AI and edge computing.

  • KubeEdge
  • KubeEdge-Sedna
  • Volcano

• Multimodal Large Model Joint Learning Algorithm: Reproduction Based on KubeEdge-Ianvs

  LFX Mentorship · Listed in the official cohort

  @CreativityHSource IssueLFX portal

  KubeEdge-Ianvs currently focuses on edge-cloud collaborative learning (training and inference) for a single modality of data. However, edge devices, such as those in autonomous vehicles, often capture multimodal data, including GPS, LIDAR, and Camera data. Single-modal learning can no longer meet the precise inference requirements of edge devices. Therefore, this project aims to integrate mainstream multimodal large model joint learning algorithms into KubeEdge-Ianvs edge-cloud collaborative learning, providing multimodal learning capabilities.

  • TensorFlow/Pytorch
  • LLMs
  • KubeEdge-Ianvs

Term 2 · Jun-Aug · 4 projects

• Iterating Enhancement for KubeEdge Dashboard

  LFX Mentorship · Listed in the official cohort

  @HongbingZhangSource IssueLFX portal

  Based on the previous release of KubeEdge, a version of the dashboard has been implemented. With the iterative updates of the backend API, the current dashboard may have several issues. In this project, we aim to iterate and update the dashboard to ensure its compatibility with the latest version of KubeEdge. Additionally, we want to refactor the dashboard using more mainstream frameworks such as Material and optimize the user experience of the dashboard.

  • KubeEdge
  • Front-end
  • nodejs

• KubeEdge Documentation Improvement

  LFX Mentorship · Listed in the official cohort

  @zhiyingfang2022Source IssueLFX portal

  Recently, we have updated the directory and structure of the community's official website documentation. We have listed some documentation improvement tasks. In this project, we would like you to have a thorough understanding of KubeEdge and complete these documentation optimization tasks to help users or developers gain a better understanding of and utilize KubeEdge effectively.

  • Kubernetes
  • KubeEdge
  • docs

• KubeEdge test cases enhancement

  LFX Mentorship · Listed in the official cohort

  @fisherxuSource IssueLFX portal

  Testing is an important task to ensure project stability, security, and other aspects. Since KubeEdge is built on top of native Kubernetes, in this project, we aim to integrate Kubernetes end-to-end (E2E) test cases into KubeEdge's CI. This integration will help ensure the native compatibility and usability of KubeEdge. Additionally, we also aim to improve the unit tests and increase the coverage of integration tests for KubeEdge.

  • Go
  • KubeEdge

• Router Manager Support HA

  LFX Mentorship · Listed in the official cohort

  @Shelley-BaoYueSource IssueLFX portal

  Users need to communicate between the cloud and the edge. For example, the cloud calls the rest interface of the edge service. In this case, the routing management function of KubeEdge can be used. Currently, routing management function of KubeEdge has some problems in the case of multiple CloudCore copies. The main problem is that when there are multiple copies of CloudCore, whether the cloud sends messages to the edge or reports the message to the cloud, it is not known which CloudCore is sent to it for processing, and there is confusion in message management in the cloud. In this project, we hope router manager can be optimized to support multi-CloudCore scenario.

  • Go
  • KubeEdge

Term 1 · Mar-May · 4 projects

• Auto Generate KubeEdge API Document

  LFX Mentorship · Listed in the official cohort

  @Shelley-BaoYueSource IssueLFX portal

  KubeEdge has introduced several custom APIs, but currently, there is no corresponding API documentation available. We would like to implement automated generation of API documentation and display it on the website documentation to assist users in quickly understanding the APIs and help developers reduce maintenance costs.

  • Go
  • Kubernetes
  • Swagger

• Image PrePull Feature Enhancement

  LFX Mentorship · Listed in the official cohort

  @Shelley-BaoYueSource IssueLFX portal

  In the latest release, KubeEdge has implemented the ability for image pre pull. However, each task execution only supports images from the same image repository currently. We hope to enhance this feature to support capabilities like overriding images and secrets, complete end-to-end tests for this feature.

  • Go
  • Kubernetes
  • KubeEdge

• Keadm Tool Enhancement

  LFX Mentorship · Listed in the official cohort

  @WillardHuSource IssueLFX portal

  Keadm(KubeEdge installation tool) now only supports configuring a subset of parameters during EdgeCore installation. We would like to support specifying parameters using --set or directly using an existing local configuration file to achieve full parameter configuration and meet the users' requirements.

  • Go
  • Kubernetes
  • KubeEdge

• Support latest version in keink and run demo on keink

  LFX Mentorship · Listed in the official cohort

  @HongbingZhangSource IssueLFX portal

  keink(represent for KubeEdge IN kind) is a project for running local KubeEdge clusters using Docker container "nodes", so developers can install a multi-node edge cluster in one node. Now we need to support the latest version installation in keink.

  • Kubernetes
  • KubeEdge
  • Go

Term 3 · Sep-Nov · 3 projects

• Add case study center in website

  LFX Mentorship · Listed in the official cohort

  @Shelley-BaoYueSource IssueLFX portal

  Now we have had many user cases in the community. However, the KubeEdge website does not have a page to display user cases. Many users lack ways to understand and learn KubeEdge implementation cases., we hope to build a case center to display them, so that more users can consult and learn.

  • JavaScript
  • HTML
  • KubeEdge

• Support latest version in keink and run demo on keink

  LFX Mentorship · Listed in the official cohort

  @fisherxuSource IssueLFX portal

  keink (KubeEdge IN kind) is a project for running local KubeEdge clusters using Docker container "nodes", so developers can install a multi-node edge cluster in one node. Now we need to support the latest version installation in keink.

  • Kubernetes
  • KubeEdge
  • Go

• Support latest version installation demo in killercoda

  LFX Mentorship · Listed in the official cohort

  @fisherxuSource IssueLFX portal

  We have created a tutorial in the interactive learning platform killercoda for KubeEdge deployment. This can give a hands-on experience of KubeEdge deployment. Now we need to support the latest version of KubeEdge and integrate example for developers.

  • Kubernetes
  • KubeEdge
  • Go

Term 1 · Mar-May · 3 projects

• Cloud-Robotic AI Benchmarking for Edge-cloud Collaborative Lifelong Learning

  LFX Mentorship · Listed in the official cohort

  @luosiqiSource IssueLFX portal

  Based on real-world datasets provided by industry members of KubeEdge SIG AI, the issue aims to build a lifelong learning benchmarking on KubeEdge-Ianvs. Namely, it aims to help all Edge AI application developers to validate and select the best-matched algorithm of lifelong learning.

  • TensorFlow/Pytorch
  • Python
  • Kubernetes

• Design and implement the KubeEdge Dashboard

  LFX Mentorship · Listed in the official cohort

  @vincentgoatSource IssueLFX portal

  Users now can use K8s API or Kubectl to talk to KubeEdge, in this project we will design and implement the KubeEdge dashboard, so users can talk to KubeEdge cluster through UI.

  • JavaScript
  • Kubernetes
  • KubeEdge
  • Html

• Re-design and implement the KubeEdge website

  LFX Mentorship · Listed in the official cohort

  @Shelley-BaoYueSource IssueLFX portal

  KubeEdge's website has been running for a few years, and now we have more customer cases and more developer courses, so this project will update KubeEdge's website, with more readable documents on the homepage, covering user cases, developer courses, etc.

  • JavaScript
  • KubeEdge
  • Html

Term 0 · 2022 Spring (Mar-May) · 4 projects

• Design and add more e2e tests especially for edge scenarios

  LFX Mentorship · Listed in the official cohort

  @wackxuSource IssueLFX portal

  We have many features for edge scenarios, as edge autonomy, reliable message transmission, etc. We need to add e2e tests for them.

  • Kubernetes
  • KubeEdge

• Move edge native k8s api interface GA

  LFX Mentorship · Listed in the official cohort

  @Shelley-BaoYueSource IssueLFX portal

  Now we have add the edge native k8s api interface, apps like operator running in edgeside can access the apiserver and obtain resources. We still need to fix bug and improve the stability.

  • Kubernetes
  • KubeEdge

• Plans for Node Group Management

  LFX Mentorship · Listed in the official cohort

  @vincentgoatSource IssueLFX portal

  In edge computing scenarios, nodes are geographically distributed. The same application may be deployed on nodes at different locations. We have plans for achieving the feature of Pod Scheduling among node groups.

  • Kubernetes
  • KubeEdge

• Updating the kubeedge docs

  LFX Mentorship · Listed in the official cohort

  @fisherxuSource IssueLFX portal

  Now we have lots of new features, we need add more docs for them.

  • Kubernetes
  • KubeEdge

Term 0 · 2021 Summer (Jun-Aug) · 2 projects

• Improve the KubeEdge website

  LFX Mentorship · Listed in the official cohort

  @kevin-wangzefengSource IssueLFX portal

  Improve the design and content of the kubeedge website.

  • JavaScript
  • KubeEdge

• Refactor the cloudStream to pass-through the request instead of parsing the web path

  LFX Mentorship · Listed in the official cohort

  @fisherxuSource IssueLFX portal

  Edgestream is used to handle the request from apiserver, then forward the request to edged through tunnel. We will find a way to pass-through the request, through the hijack stuff, instead of parsing the web path manually.

  • Go
  • Kubernetes
  • KubeEdge

Term 0 · 2021 Spring (Mar-May) · 3 projects

• Design more tests for specific scenarios of edge computing

  LFX Mentorship · Listed in the official cohort

  Source IssueLFX portal

  We need to do some designs for adding more tests especially for the specific scenarios of edge computing, eg: Application migration when the network is disconnected System stability when the network is unstable Run large-scale cluster tests periodically

  • Go
  • KubeEdge

• Integration and verification of third-party CNI/CSI based on the edge side list-watch

  LFX Mentorship · Listed in the official cohort

  @GsssCSource IssueLFX portal

  We need to integration and verification of third-party CNI/CSI for the edge applications.

  • Go
  • KubeEdge

• Support multi-instance high availability cloudcore for large-scale cluster

  LFX Mentorship · Listed in the official cohort

  Source IssueLFX portal

  Cloudcore is the core component of kubeegde in cloud, which is responsible for sending resources of cloud to the edge. Now the cloudcore is running in leader/follower mode, only one instance can run at the same time. For the more larger scale cluster, we need to support multi-instance high availability for cloudcore.

  • Go
  • KubeEdge

Term 0 · 2020 Q3–Q4 · 3 projects

• Add EdgeGateway as the ingress gateway on Edge

  CommunityBridge · Listed in the official cohort

  @kevin-wangzefengSource IssueCommunityBridge

  Add the EdgeGateway as the ingress gateway on Edge to dispatch request to multiple backend pods on the Edge side.

  • KubeEdge
  • Kubernetes
  • Go

• Support list-watch from edgecore for applications on the edge

  CommunityBridge · Listed in the official cohort

  @kevin-wangzefengSource IssueCommunityBridge

  Some applications running on the edge side need to connect to the k8s master through list-watch interface, but on the edge it cannot directly connect to the k8s master. So we need forward the list-watch requests to the k8s master through the cloud-side channel.

  • KubeEdge
  • Kubernetes
  • Go

• Use device API both on cloud and edge

  CommunityBridge · Listed in the official cohort

  @kevin-wangzefengSource IssueCommunityBridge

  Now we have defined device API in the cloud side, but the edgeside(edgecore and mapper) still don't use the API, we should refactor edgeside and use same API on cloud and edge to reduce complexity.

  • KubeEdge
  • Kubernetes
  • Go

Term 0 · 2020 Q2 · 2 projects

• Add certificate rotation for edge node

  CommunityBridge · Listed in the official cohort

  @kevin-wangzefengSource CommunityBridge

  At present, the edge nodes of KubeEdge only apply for certificates when they first started. And we need to automatically generate a new key and request a new certificate from the cloudcore periodically.

  • Go
  • Kubernetes

• Support metrics-server in cloud

  CommunityBridge · Listed in the official cohort

  @kevin-wangzefengSource CommunityBridge

  Now we have integrated cadvisor in edge, and it also needs to integrate metrics-server in the cloud to collect monitoring data of edge nodes.

  • Go
  • Kubernetes