Skip to content

Baetyl-cloud v2.2.0 release
Compare
Choose a tag to compare
@xiejingru xiejingru released this 28 Jun 06:59
· 254 commits to master since this release
v2.2.0
efe124f
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.

在过去的一段时间里，Baetyl在众多的活跃贡献者的努力下，实现了更多有挑战性的功能。这些新的特性依然本持着云原生的理念，继续朝着构建开放、安全、可扩展、可控制的智能边缘计算平台的方向前进。
具体来说，相较于之前发布的Baetyl v2.1.1版本，v2.2.0 版本提供的新特性包括：

  • 新增了对EdgeX Foundry的支持
  • 新增了对支持边缘集群环境的所需的API定义
  • 新增了对DaemonSet负载类型应用的支持
  • 新增了对已部署应用的远程调试、远程日志查看的API定义
  • 新增了对GPU 监控及共享功能的API定义
  • 更多的官方模块

以上的特性在边缘计算场景下都是需求比较急迫且具有很高实用价值的功能。
关于对EdgeX Foundry的支持。Baetyl v2.2.0版本已经完成对开源计算框架EdgeX Foundry的兼容，通过baetyl的云端管理套件，你可以向边缘侧下发完整的EdgeX 14个服务，下发的EdgeX服务将由baetyl提交部署并统一监测信息与云端进行信息同步。
关于对边缘集群的支持。在工业物联网场景中，经常会有大量工控盒子构成一个边缘集群的场景。Baetyl定义了开放的多集群管理API，通过实现这些API，可以把整个集群反映在云端控制台上，你可以方便的将应用部署到定义的集群中，更进一步的你可以指定部署到某些边缘子节点上。
关于对DaemonSet负载类型应用的支持。在支持了集群的背景下，类似于针对集群中每个节点状态监控的功能就需要一种新的负载方式来支持部署，所以我们也对DaemonSet进行了支持，通过这种负载类型，可以将服务在匹配到的每个集群中的节点上启动一个副本，并会随着新节点的增删自动调整。
关于远程调试、远程日志查看功能。为方便对边缘设备进行调试或日志查看操作，我们建立了开放的远程调试API，能够在未来与多种云端控制系统打通。
关于对GPU的支持。对GPU的支持主要包括两方面，一是对GPU的使用监控，二是对GPU共享的支持。通过GPU监控模块，baetyl-core可以实时获取GPU当前显存使用量、温度、能耗等信息。通过GPU共享功能，多个应用可以共享使用设备的GPU资源。目前已经完成了GPU支持接口的定义，仅需在端侧提供一个包含GPU share功能的模块即可使用。
在边缘侧，也提供了更多的官方系统模块：

  1. baetyl-init: 负责激活边缘节点到云端，初始化并守护 baetyl-core，任务完成后会持续上报并同步core状态；
  2. baetyl-rule: 可以实现 baetyl 框架端侧的消息流转，在 baetyl-broker (端侧消息中心)、函数服务、Iot Hub (云端 mqtt broker) 进行消息交换。

在这些新特性之外，还提供了很多其他功能细节上的优化和机制上的完善，比如安装过程的优化、系统应用可根据需要选择配置、定义了事务执行接口、任务队列接口等。
以上新功能会随着 Baetyl 2.2.0 的正式发布立即可用。

Baetyl has now achieved more advanced functions with the efforts of many active contributors. These new features still adhere to the concept of Cloud Native, and continue to move towards building an open, secure, scalable, and controllable intelligent edge computing platform.
Specifically, compared to the previously released Baetyl v2.1.1 version, the new features provided by the v2.2.0 version include:

  • Support for working with EdgeX Foundry
  • New API definition needed to support edge cluster environment
  • Support for DaemonSet load type applications
  • New API definitions for remote debugging and remote log viewing of deployed applications
  • New API definitions for GPU monitoring and sharing functions
  • More official modules

The above features are functions that are in urgent need and have high practical value in edge computing scenarios.
About support for EdgeX Foundry. The Baetyl v2.2.0 has completed compatibility with the open-source computing framework EdgeX Foundry. Through baetyl's remote management suite aka. Baetyl-cloud, users can deliver all 14 EdgeX services to the edge. The delivered EdgeX services will be submitted by Baetyl to local Kubernetes cluster for deployment and monitoring synchronized to the cloud.
About support for edge clusters. In the industrial IoT scenario, there are often many industrial control boxes to form an edge cluster scenario. Baetyl defines open multi-cluster management APIs. By implementing these APIs, the entire cluster can be reflected on the cloud console. Users can easily deploy applications to defined edge clusters and specify edge node affinity within cluster.
About support for DaemonSet load type applications. In the context of supporting clusters, a new load method is needed to support deployment like the function of monitoring the status of each node in the cluster, so baetyl also support DaemonSet. Through this load type, the service can be a single replica launched on every node in the matched cluster, and it will be automatically scaled to new added nodes and vice-versa.
About remote debugging and remote log viewing functions. To facilitate debugging or log viewing operations on edge devices, baetyl have established an open remote debugging API that can connect with multiple cloud control systems in the future.
About GPU support. The support for the GPU mainly includes two aspects, one is the monitoring of the use of the GPU, and the other is the support for the GPU sharing. Through the GPU monitoring module, baetyl-core can obtain the current GPU memory usage, temperature, energy consumption and other information in real time. With the GPU sharing function, multiple applications can share the GPU resources of the device. At present, the definition of the GPU support interface has been completed, and only a module containing the GPU share function needs to be provided on the end side to use.
There will be more official system modules are also provided:

  1. baetyl-init: Responsible for activating edge nodes to the cloud, initializing and guarding baetyl-core, after the task is completed, it will continue to report and synchronize the core state.
  2. baetyl-rule: can realize the message flow of baetyl framework end-side, and exchange messages in baetyl-broker (end-side message center), function service, IoT Hub (cloud MQTT broker).

In addition to these new features, it also provides many other functional details optimization and mechanism improvement, such as the optimization of the installation process, the system application can be configured according to needs, and the transaction execution interface and task queue interface are defined.

All these new features will be available immediately with the release of Baetyl 2.2.0.

Assets 2
Loading