Kubernetes Native Edge Computing Framework (project under CNCF)
APACHE-2.0 License
Bot releases are hidden (Show)
Published by Shelley-BaoYue 6 months ago
See CHANGELOG-1.16.md for details.
Published by kevin-wangzefeng almost 2 years ago
See CHANGELOG-1.13.md for details.
Published by vincentgoat almost 2 years ago
See CHANGELOG-1.12.md for details.
Published by kevin-wangzefeng over 2 years ago
See CHANGELOG-1.9.md for details.
Published by kevin-wangzefeng over 2 years ago
See CHANGELOG-1.10.md for details.
Published by kevin-wangzefeng over 2 years ago
See CHANGELOG-1.11.md for details.
Published by fisherxu over 3 years ago
See CHANGELOG-1.7.md for details.
Published by kevin-wangzefeng about 4 years ago
See CHANGELOG-1.4.md for details.
Published by fisherxu about 4 years ago
Published by fisherxu over 4 years ago
See CHANGELOG-1.3.md for details.
Published by kevin-wangzefeng over 4 years ago
See CHANGELOG-1.3.md for details.
Published by kevin-wangzefeng over 4 years ago
Published by kevin-wangzefeng over 4 years ago
See CHANGELOG-1.2.md for details.
Published by fisherxu about 5 years ago
See CHANGELOG-1.1.md for details.
Published by fisherxu about 5 years ago
Fix edge_core cannot connect to edgecontroller after disconnecting once (#870, @shouhong)
Raspberry Pi3/4 cross build (#903, @subpathdev)
Upgrade to Kubernetes v1.15 (#941, @edisonxiang)
Add New Feature: support dockershim in edged. (#829, @arcanique)
Use go mod (#947, @subpathdev)
Rename device dir to mappers which places mappers (#966, @fisherxu)
New feature: L4 Proxy support in edgemesh (#970, @arcanique)
Initialize feature lifecycle doc (#850, @kevin-wangzefeng)
Add admission webhook for validate device CRD (#984, @chendave)
Add in-tree csi plugin implementations (#1047, @edisonxiang)
Add csi driver from kubeedge (#1059, @edisonxiang)
Published by sids-b over 5 years ago
Edge Mesh
This feature aims to support service mesh capabilities on the edge to support microservice communication cross cloud and edge. In v1.0.0 release, pod-to-pod communication on the same edge node or across
edge nodes in the same subnet is supported.
CRI support
This feature enables edged to communicate with a CRI-compliant runtime to manage containers running on resource constrained edge nodes. Support for containerd is tested.
Quic protocol support
In order to enhance cloud and edge communication efficiency, communication between the edge and the cloud is now also supported via QUIC, a UDP-based protocol. CloudHub supports both Websocket and QUIC protocol access at the same time. The edgehub can choose one of the protocols to access the cloudhub.
Modbus Mapper
Modbus Mapper is an application that is used to connect and control devices that use Modbus(RTU/TCP) as a communication protocol. The user is required to provide the mapper with the information required to control their device through the dpl configuration file. These can be changed at runtime by updating configmap.
Edge Site
Edge Site enables to run a standalone Kubernetes cluster at the edge along with KubeEdge to get full control and improve the offline scheduling capability. A Kubernetes cluster is deployed at edge location including the management control plane. For the management control plane to manage a reasonable scale of edge worker nodes, the host master node needs to have sufficient resources.
Reliable message delivery is missing between cloud and edge.
Installer currently doesn't support installation of containerd, cni-plugins.
Port mapping is not supported in CRI
Pods created using CRI on deletion remain in terminating status. (#755, @gpinaik)
fix bug: replace value in key, value by expected (#759, @RicardoZPHuang)
update edge_core version to reflect the vendor k8s and kubeedge version (#761, @sids-b)
Edgesite documentation (#737, @naveensriram)
Changes made to Installer to support CRI (#795, @srivatsav123)
Published by sids-b over 5 years ago
Edge Mesh
This feature aims to support service mesh capabilities on the edge to support microservice communication cross cloud and edge. In v1.0 release, pod-to-pod communication on the same edge node or across
edge nodes in the same subnet is possible.
CRI support
This feature enables edged to communicate with a CRI-compliant runtime to manage containers running on resource constrained edge nodes. Support for containerd is tested.
Quic protocol support
In order to enhance cloud and edge communication efficiency, communication between the edge and the cloud is now also supported via QUIC, a UDP-based protocol. CloudHub supports both websocket and QUIC protocol access at the same time. The edgehub can choose one of the protocols to access the cloudhub.
Edge Site
Edge Site enables to run a standalone Kubernetes cluster at the edge along with KubeEdge to get full control and improve the offline scheduling capability.
Modbus Mapper
Modbus Mapper is an application that is used to connect and control devices that use Modbus(RTU/TCP) as a communication protocol.
Published by sids-b over 5 years ago
Published by rohitsardesai83 over 5 years ago
Device Management using CRD’s : This feature mainly provides api’s for managing devices from cloud and synchronize the device updates between cloud and edge. DeviceController is the module in cloud which is responsible for watching on devices and devicemodels and syncing the data to edge. DeviceController is also responsible for receiving the device-twin updates sent from edge and making the changes in cloud(apiserver). DeviceController also creates a configmap for each node and adds the deviceInstances, protocols, properties and propertyVisitors in it. This configmap can be consumed by any application like mapper and used to configure/control devices.
Bluetooth Mapper : Bluetooth Mapper is an application that is used to connect and control devices that use bluetooth as a communication protocol. It consumes the configmap generated by device-controller and reuires a configuration file for information not present in configmap. Configurations can be changed at runtime by providing the input through the MQTT broker. It can run schedule based actions and converts data received from devices into edge_core expected format.
Performance Test Framework : Benchmarking framework to measure the performance against the following Service Level Objectives:
Latency: time cost from the moment when the server gets the request to last byte of response sent to the users.
Throughput: measure how many requests can be served within given time.
Scalability: potential scaling capacity (including number of nodes, pods, devices etc. at the edge) under different load conditions.
CPU Usage: measure the cpu usage of KubeEdge under different load conditions.
Memory Usage: measure the memory usage of KubeEdge under different load conditions.
Api Validation for device controller is missing.
UT/e2e coverage should be improved for cloud part.
Partitioning of device configmap.
Running edged and k8s master on same VM kills k8s master pods.
CRI support in edged.
EdgeMesh for edge to edge communication.
Evaluation and enhancement of cloud-edge websocket connection using QUIC.
EdgeSite support for KubeEdge.
Published by rohitsardesai83 over 5 years ago
https://github.com/kubeedge/kubeedge/pull/354
https://github.com/kubeedge/kubeedge/pull/328