LF Edge will be on-site at ONS on September 23-25 in Antwerp, Belgium. We invite you to join us in a Mini-Summit (details to come) and any one of the edge talks led by member thought leaders. Please see below and add these to your schedule today.
Monday, September 23:
4:15 – 4:45 pm: Integrate ROS into EdgeX – Startover of Edge Computing Systems – Tiejun Chen, VMWare
ROS – Robot Operating System is a flexible framework for writing robot software. It is a set of tools, libraries, and conventions of creating complex and robust robot behavior across a wide variety of robotic platforms. Essentially it is a middleware over Linux but more importantly, it is broadly adopted across many industrial automated platforms and even automotive systems. In this presentation, Tiejun will review if we can integrate ROS into EdgeX Foundry to build a unified edge system solution like 1: Architecture; 2: EdgeX with ROS device nodes; 3. a basic machine learning framework with ROS library like opencv; 4: Over light weight preempt-rt Linux; 5. vision of edge computing cluster. Add this to your schedule here.
5 – 5:30 pm: Panel Discussion: Akraino – Balaji Ethirajulu, Ericsson, Tapio Tallgren, Nokia, Tina Tsou, Arm, Kandan Kathirvel, AT&T, and Sukhdev Kapur, Juniper Networks
The panelists will discuss LF Edge and how the Akraino Edge Stack fits in to overall umbrella. They will also present Akraino architecture, use cases, business benefits, and blueprints. Add this to your schedule here.
Tuesday, September 24:
11-11:30 am: Your path to Edge Computing – Kandan Kathirvel, AT&T, and Tina Tsou, Arm
The Akraino community has experienced extremely rapid growth over the past year in membership and community activity. In fact, Akraino announced the availability of its release in June with additional blueprints. An Akraino Blueprint is not just a diagram; it’s real code that brings everything together so users can download and deploy the edge stack in their own environment to address a specific edge use case. In this talk, attendees will learn about R1 blueprints and its use, R2 goals and how to engage and contribute to the Akraino Community. Add it to your schedule here.
1:50-2:20 pm: Project EVE: Enterprise/On-prem Edge Virtualization – Erik Nordmark, Zededa
The enterprise edge is different from traditional data center environments:
– Physical distribution and resulting physical and network security
– Service lifetime of 7 or more years
– Security requirements for cyber-physical systems
Yet system designers want to follow cloud native and devops approaches for deploying apps at the edge. This is why we created the Edge Virtualization Engine, or Project EVE. This talk covers requirements of the enterprise edge environment, overview of Project EVE , Zero Trust, Zero Touch for the enterprise edge, providing freedom of any app, any network (technology), and any hardware and a brief demo. Add it to your schedule here.
2:35-3:05 pm: Android at the Edge – Melissa Evers-Hood, Intel
As cloud architectures continues to gain prominence, it is extending from cloud all the way to the device edge. There are many benefits to device and edge in cloud software infrastructure, but there are also differentiated challenges- a UI for example! The prevalence of this migration to device with cloud software is seen across LF Edge projects. Many folks are looking to bring the exceptionally popular device operating system, Android, to these cloud architectures in either VM or Containerized implementations. This drive for workload consolidation and convergence presents enormous opportunities ( >2M app ecosystem, robust tooling, Google services) as well as challenges (certification, performance tuning, security). We will review LF Edge projects and pathfinding development done with AOSP and show demos of Android in a variety of scenarios. Add this to your schedule here.
3:20-3:50: Akraino KNI: Project Update – Ricardo Noriega & Yolanda Robla Mota, Red Hat
Blueprints in the Kubernetes-Native Infrastructure Blueprint Family leverage the best-practices and tools from the Kubernetes community to declaratively manage edge computing stacks at scale and with a consistent, uniform user experience from the infrastructure up to the services and from developer environments to production environments on bare metal or on public cloud. This session will make an introduction of the Akraino KNI blueprint family, and show what has been released in Akraino R1. All blueprints in this family share the following characteristics: K8s Machine API (declarative API to manage a configure the infrastructure of a cluster), Operator Framework (automated and secure lifecycle of applications running on the edge stack), Kubernetes-native workloads (allowing the mix of CNFs and VM-based workloads via Kubevirt). Add this to your schedule here.
4:20-4:50 pm: Addressing IoT Protocol Soup – Leveraging a Highly Interoperable, Flexible and Open Source Edge Platform – Jason Shepherd, Dell Tech
Today’s IoT/edge landscape often requires developers to connect a variety of sensors and devices speaking dozens of protocols into an IoT platform. The IoT platforms, after ingesting data from the sensors and devices, must prepare the data for a variety of applications, cloud systems, and analytic engines. Developers must negotiate a diversity of endpoint protocols, data formats, and special data needs. Thus, IoT platforms today and for the foreseeable future are dual transformation engines – converting OT data into IT data and then arranging it for enterprise systems and applications. In this session, learn how EdgeX Foundry provides a scalable and flexible platform that addresses the dual transformation needs of the edge. Add it to your schedule here.
Wednesday, September 25:
9-9:25 am: Keynote Panel – Unifying the Edge – IoT, Enterprise, Telecom and Cloud – Melissa Evers-Hood, Intel; Erik Nordmark, ZEDEDA; Jason Shepherd, Dell; moderated by Will Townsend, Moor Insights & Strategy
Add it to your schedule here.
12-12:30 pm: Multi-Access, Multi-Edge Access with Akraino – Tapio Tallgren, Nokia
The ETSI Multi-Access Edge Computing (MEC) started with the idea that a base station can host a server. The benefits of being close are low latency — critical for AR/VR applications — and processing information close to the producer so that all does not need to be transmitted. The Akraino MEC API Framework project creates tools that translate the ETSI APIs to programming language interfaces, and provides a repository where applications can find available APIs. The Akraino uMEC project is an ultra-small MEC platform that can be deployed in a lightpole in a Smart City, along a Smart Highway, or in a liftfork in a Smart Factory. We tested the uMEC concept in a hackathon in Finland, with the task to create new applications for smart city denizens. We are planning to create a demo platform next that would allow the participants to create applications that will use real MEC APIs. Add this to your schedule here.
1:50-2:20 pm: Panel Discussion: Harnessing the Kubernetes (K8s) Ecosystem for Edge Automation & Computing – Ramki Krishnan, VMware, Davide Cherubini, Vodafone, Mike Elliott, Amdocsm Srinivasa Rao Addepalli, Intel, Catherine Lefevre, AT&T
Edge computing industry is increasingly using cloud technologies for seamless migration of workloads across edges and clouds. For seamless mobility workloads, K8s is a key requirement for all CSPs. Also, K8s is being thought to be good infrastructure orchestrator for all deployment types (VMs, containers and functions). As an example, collection/learning/inference components of a typical analytics app are distributed across edges and clouds. Many applications, including ONAP distributed analytics work assumes that the edges and clouds are K8s enabled. This panel discusses existing work and novel ways of realizing a consistent app lifecycle management across distributed clouds using the extensibility of K8s ecosystem. This work is currently happening in ONAP as part of the Edge Automation effort and we see this impactful to other open source efforts such as Akraino, K8s Edge WG etc. Add this to your schedule here.
2:35-3:05 pm: Enabling DPDK in Akraino and Airship for 5G networks – Gerog Kunz, Ericsson and Cheng Li, Intel
The Akraino project is defining and implementing reference software stacks to enable edge use cases. In case of 5G networks, network performance is key. For this reason, we have been working on a performance optimized software stack in Akraino by adding support for OpenVswitch with DPDK to the Network Cloud Unicycle blueprint. To this end, we introduced DPDK support to the underlying Airship and OpenStack-Helm projects.
In this presentation, we will talk about our motivation for enabling DPDK support in the Network Cloud blueprint. We will introduce the high-level design of the software stack and provide an overview of how we leverage it for 5G vRAN deployments. Finally, we will dive into the technical implementation in the Airship and OpenStack-Helm projects and showcase how we achieved and leveraged cross-project collaboration. Add this to your schedule here.
3:20-3:50 pm: Edge Computing Offloading – Juraj Veverka, PANTHEON.tech
The promise of 5G: low latency and high bandwidth. In this talk we will discuss how Radio Access Network could be used to offload tasks from the end-user device to the edge. We will demonstrate a client application (mobile device) that can execute compute-intensive tasks on an edge compute node. Project EVE has been used on a RAN site to host specialized container application, based on TensorFlow and pre-trained models, to classify image content, created on client devices. Working example with source code will be presented.
3:20-3:50 pm: ONAP Orchestrated Edge and Cloud Collaboration in China Unicom – Chang Cao, China Unicom and Derek Chen, Huawei Technologies
The business scenario of this joint demonstration is for industrial Internet, especially in the smart campus extracts video or image abstracts from the edge side and processes them on the cloud to prevent massive video data from being transmitted to the cloud and occupy limited resources. Based on the automatic orchestration and scheduling capabilities of ONAP, this topic implements cloud and edge collaboration and computing sinking. Based on the Akraino Eliot project capability, this demonstration extends the ONAP application management capability to the edge and solves the challenges faced by users in the current intelligent edge domain. ONAP with Akraino implements orchestration of edge to ensure the low-latency experience for end-users and no need to upload the huge video and pictures data to save bandwidth. Add this to your schedule here.
5-5:30 pm: Use Case – The Policeman’s Camera is on Duty with Akraino and EdgeX Foundry Technology Embedded – Xuan Jia, China Mobile and Tuejun Chen, VMware
When a policeman is walking in the street, there will be a small camera which hangs on his breast pocket. It could record the video and send the stream into the edge server, then we do face recognition in the edge server. At the last, policeman’s cellphone will get to know when suspect shows up. In this presentation, we will share how we integrate Akraino, which provides the basic edge cluster, and Edgex Foundry, which provides the camera capability. In the edge cloud, we deploy china mobile edge computing platform to do AI analysis. This platform is based on Kubernetes, it provides GPU and AI capability to fulfill this use case’s requirement. Add this to your schedule here.
