Akraino KubeEdge blueprint project builds real world Edge computing user cases based on KubeEdge. The first case of this project is ML Offloading. With Artificial Intelligence democratized into mobile phones, the computing resource limitation of mobile devices is increasingly a concern. More and more apps running on mobile phones have to offload ML related operations, starting with the public cloud. Offloading to edge is a natural thought for latency and data security considerations.
Traditional IoT Gateways often serve as a proxy between devices and a Cloud where all control operations take place. It mainly requires time-series dataflow processing.
Coordination and control operations require a stateful model of the observed environment in terms of high-level business objects (“as is” and “to be”), derived from low-level sensor data. Eclipse Ditto or other Digital Twins As a Service systems could be used on the Cloud to store and query a state of devices and assets and decouple it from implementation details of drivers and protocols.
2019 marked the 50th anniversary of the first moon landing. Decades later, the Apollo program still stands as one of humanity's most impressive technical achievements. To think that we sent people in the void of space using a computer running at 1.024MHz and around 76K of memory!
Edge computing is emerging, and many open source organizations have edge projects, such as Linux Foundation Edge, CNCF, and Eclipse as well. There should be more cooperation between open source community to provider more efficient solution for customers. As we know that many excellent IoT projects in Eclipse, therefore, we will introduce the Eclipse IoT Packages integration with CNCF KubeEdge, which a cloud native project in edge.
Lately, many IoT use cases evolve toward the Edge architectures. We'll start the session by exploring these use cases and accommodating architectures, trying to summarize all pros and cons of adopting them.
Eclipse IoT and Edge communities already provide many projects that are well established and solve many problems in their respectable domains. But there is an opportunity for more integration that would provide a better off-the-shelf IoT Edge experience.
Edge computing is all the rage right now. Some people even say it is more important for the future of IT than cloud computing itself. Is this real, or is this just hype? In theory, edge computing helps solve the challenges of bandwidth, latency, resiliency, and data sovereignty. Those benefits, however, will be hard to realize in the real world if you rely on a platform built by people just riding a wave. You need a real platform with real users, rooted in lessons from the trenches.
This session will walk through several reference architectures that solve common edge and IoT situations. The audience will gain a solid understanding of how to place many popular Eclipse IoT technologies in a software stack and form a complete set of solutions.
With the raise of Edge and Fog Computing there is a growing need to maintain data locally, or at least as close as possible to where it is produced, while at the same time making it accessible globally. Yet, existing storage technologies are geared toward cloud-centric deployments, in which a pool of servers is used to shard and replicate the data. As a result there is a mismatch between the storage needs of Edge and Fog Computing applications and available technologies.
The recent trend of cloud-native computing and microservices architectures have taken developers by the storm. But what if we want to move our services closer to devices and users that generate and use data? Enter the Edge computing, trying to extend cloud-native computing beyond the centralized data centers. In this session, we'll try to get you interested in this topic, talking about why and how you can start introducing Edge computing in your IoT projects.