Open Source Summit + ELC Europe 2019

Edge devices are slated to exceed 20 billion by 2020. For IoT to succeed in these devices, the currently fragmented edge market needs to be able to work together to identify and protect against problematic security vulnerabilities and advance a common, constructive vision for the future of the industry.

The recently launched LF Edge is part of this solution. It is an umbrella organization to establish an open, interoperable framework for edge computing independent of hardware, silicon, cloud, or operating system, will help ensure greater harmonization to accelerate deployment and bridge the gaps in the open source technologies to support the emerging Edge use cases.

In this panel, attendees will learn:
1. Edge use cases being addressed to satisfy the industry needs
2. Collaboration between the LF Edge Projects and scope of each project
3. How to engage and contribute to each project

EdgeX is an open source hardware-agnostic framework for IoT and Edge Computing. It is composed of a series microservices that allow developers to plug and play from the growing ecosystem of third-party components.

Until now, the only way to interact with EdgeX was to consume the different microservices' APIs through curl requests or Postman and through complicated developer scripts. As a newcomer to the project, this represents a high barrier to entry. With the EdgeX CLI, we hope to lower the high barrier to entry and improve developer productivity by offering a single user-friendly point of access. With the tool, users can perform actions on devices, initialize a new EdgeX instance, connect to a remote instance among other things.

In this talk, we share our experience creating a CLI tool in Go for the EdgeX project. We share best practices and the anatomy of a good CLI by looking at other popular CLI tools. We also discuss libraries that can help you jump start your CLI development.

The technology evolution drives a significant continuous increase in the total number of connected devices and the overall generated traffic. This pushes network service providers, like ISPs and Telcos, to carefully consider their technology choices when designing their next generation architectures. Unfortunately, the current generation of container networking solutions doesn't fit very well to that preventing industries with more sophisticated network demands from adopting the new cloud-native paradigm. The newly adopted CNCF Sandbox project, Network Service Mesh, offers a potential solution. In this talk we'll introduce the novel approach Network Service Mesh uses to solve complicated L2/L3 challenges in Kubernetes. We will also discuss use cases and provide a set of examples of this solution.

Edge and IoT are coming to us fast in the industry after the clouds and mobiles, while people are still questioning about what's the edge, and what's the best way to compile and run applications on the edge? Different perspectives come from different standing points. Developers in the clouds see and advocate no much difference on the edge compared with the cloud, and veterans on embedded systems have totally different view points, looking from the other end of the spectrum. With the recent trend of convergence for new IoT and legacy industry control systems, we see they blur together quickly.

This presentation is about the elaboration of edge native applications, how they are different from legacy applications and cloud native applications, and how to build edge native applications, well integrated with the clouds. The recommendation toolkit is relative to Linux, container, EdgeX Foundry, and a few other OSS projects.

IoT Edge processing as an architectural component emerged to meet the demands of reducing network bandwidth requirements and lowering response latency. Microservices as a software paradigm evolved to facilitate agile development of large-scale systems. IoT applications are sensor data triggered, matching well the event-driven paradigm of Microservices. However, Edge resources (CPU/RAM/storage/network capacity and availability) can range from marginal to significant depending on the use case. How does this affect Edge Microservice design? There are ramifications from number and payload size of inter-service calls, data caching, database accesses, network connectivity and capacity, system scalability, and consistency. We illustrate their effects in the context of EdgeX, an open source Edge IoT project, and share performance results from the Delhi, Edinburgh, and Fuji releases. We conclude with best practices, customizable deployments, and the need for monitoring and management.

The problem: Widespread fragmentation and the complexity of the current IoT landscape are stalling the growth of the Industrial IoT market and making it difficult for partners to work together. The overwhelming number of software platforms, applications, hardware and connectivity standards is creating paralysis among businesses that are afraid of making the wrong choice and being locked into a solution.

The Solution: LF Edge, an umbrella organization under the Linux Foundation, aims to create an open interoperable framework for edge computing independent of hardware, silicon, cloud, or operating system. LF Edge helps ensure greater harmonization to accelerate deployment and bridge the gaps in the open source technologies to support the emerging Edge use cases. Attendees will learn more about the path to edge computing with Akraino Edge Stack, EdgeX Foundry, Home Edge, Project EVE and the Open Glossary of Edge Computing.

IoT sensor readings are ingested by an edge platform in many different protocol formats. Using EdgeX Foundry (an open source edge platform) this disparity can be normalized into a common representation. Even so, this data may not be secure or tracked until it gets to the enterprise. Can you securely store and track IoT sensor data at the limited-compute-resource edge with distributed ledger? Yes! In this session, come learn how to secure and track data entering the IoT edge using distributed ledger technology. Specifically, see how data brought into the EdgeX Foundry can be signed and stored on an IPFS distributed file system and then an immutable pointer to the data can be placed into a distributed ledger along with any necessary identity/wallet information. This solution helps minimize storage needs of the ledger, provides an extra layer of security to your IoT solutions and can pave the way for eventual data monetization via querying/reporting capabilities

Machine Learning (ML) has been happening only in Cloud and some inference on Edge. But there's quite few ML happening in IoT where Linux cannot fit. This area is called TinyML. In order to democratize ML in IoT, I propose TinyML as-a-Service, which would bring ML onto microcontrollers with 3 enablers, (1) Unikernel, (2) ML compiler and (3) CoAP. In this session, I'll explain our end-to-end experiment of how we tackled to bring ML onto Edge & microcontrollers, show demo and also discuss future possibilites. The audience are not expected much. The following links may be helpful if you discuss this session deeper but I'll explain them, anyway;) Enjoy!! #tinymlaas
 
Resources:

• Unikernel
• MirageOS
• Ocaml
• Owl
• LwAE
• TinyML
• Tensorflow Lite for micro
• LwM2M
• CoAP

Any further discussion would be really apprecaited;)