Download presentation
Presentation is loading. Please wait.
Published byChantal Tiedeman Modified over 5 years ago
1
A fast way to deploy IoT application based on EdgeX Foundry
Ximing Xu
2
We manage smart devices and machine data
Smart devices management ThunderUEM Access and Identity Management ThunderIAM RealTime Data InstantD Products Solutions OneIoT Manage kinds of IoT devices used in industrial scenario, and realtime data generated by these devices. OneSpace Provide a controllable digital working space across PC and mobile environment. +
3
Strategic investors AirWatch
A global leader in cloud infrastructure and business mobility, helps customers accelerate their digital transformation. With 2016 revenue of $7.09 billion, VMware is headquartered in Palo Alto, CA and has over 500,000 customers and 75,000 partners worldwide. Airwatch of VMware is the clear leader for EMM globally recognized by industry analytics and key customers. AirWatch Comprehensive Portfolio Innovate Faster Market Leader Built to Scale Open Ecosystem Security World leading smart device operating system and platform technology provider Partnering with Qualcomm, Intel, ARM, Microsoft and have joint ventures with Qualcomm, ARM, VMware etc. Core competencies include protocol stack, deep learning, computer graphics, OS optimization, etc on Android, Linux, Windows and HTML5
4
The challenges for industry IoT application
Too many I/O types and communication protocols. - IO types: RS485, RS232, CAN, Ethernet, Zigbee, LoRa, GPRS, NB-IoT …… - Protocols: Modbus RTU, Modbus TCP, OPC-UA, TCP/IP, Http, MQTT, CoAP, PPI, Profibus, CANopen, KNX …… Different data processing requirements for different devices. Difficult to deploy basing on existing industry infrastructure. IoT Maintenance costs are extremely huge.
5
Existing solutions for the challenges IoT
Smart device instruments Smart gateway SCADA system 5
6
Why EdgeX Foundry will be new trend in industry IoT?
Easy-using: focus on application itself plenty of components Flexible: assemble micro-services at will. services loosely coupled. support both single server deployment and complex fog env. Easy-using Flexible Free: totally open source Independent: run on multiple platforms. support multiple OS. support multiple programming languages. Free Independent
7
We will contribute instantEdge to EdgeX Foundry
Our goal is to make it fast to deploy IoT application. Stable and scalable streaming data processing runtime. Rich computing nodes including data source,processor, sink, predict model, external data. UI editor to assemble data processing flow. Online monitor of data processing and data visualization. 7 7
8
InstantEdge architecture overview
External Source Stream Processor Export Target Service Layer Node Layer Source ZoreMQ Consumers Kafka Consumers TCP channels Https services DB connectors Service callers Schedule service Message queue Process Sink Message queue Filter Compare Function Http Sender EdgeX gateway EdgeX gateway biz system Window cache Sort MQTT publisher Devices external data Discard Merge DB sinker Archived data Data getter Alert Map Align Msg sender Predict Job Editor IOT platform Device data Topology Generator Stream Engine JS executor Window caching Data aggregator Caching service sending Mlib …… Http pusher MQTT sender DB updater Msg sender External DB
9
Embedded with EdgeX Foundry
Flow-based GUI: Flow Edit UI is the main interface for users to create and edit computing flows in a visual way Deploy Engine is responsible for deploying certain flows to a designated “edge” Stream Monitor collects state information from running streams and displays to users Flow runtime: A “Runtime” is the core for scheduling in-memory computing jobs Flow runtime will register itself as an export client and receive events from Distribution service. Flow runtime will process each events according to computing nodes。 Compute node types include filter, transform, function, split, window, join, sort, notify, export, external data…… Flow Edit UI Deploy Engine Stream Monitor GUI Runtime Export Services Distribution Registry Support Services Flow runtime logging Core Services Core Data Command Meta Data Data Services Modbus SNMP ……
10
Example 1:real-time computing and monitoring
EdgeX Gateway temperature sensor1 msg router add prosperities: msg.device_sn, msg.device_key… msg.temp= msg.t/10 group by device_sn: cache latest status Data Source map temp_sensor_profile1 humidity_sensor_profile2 function msg.hum= msg.h/10 cache cache historian data in one day latest,max,ave,min… aggregate msg.temp>40.0 trigger msg.hum<10 ,sms alert send to cloud sink humidity sensor1 temperature sensor2 humidity sensor2
11
Example 2: daily oil yields by each pump unit
EdgeX Gateway pump1 pump2 pump3 pump4 msg.device_id=1 ->msg.device_sn=‘pumpA’ msg.device_id=2 ->msg.device_sn=‘pumpB’ msg.device_id=3 ->msg.device_sn=‘pumpC’ msg.device_id=4 ->msg.device_sn=‘pumpD’ …… map function group by device_sn: cache historian data in one day cache msg.daily_yield =sum(vol*oil_percent)/ sum(delt_time)*1440 send to cloud sink Data Source
12
Example 3: prediction of defective rate
EdgeX Gateway conveyer belt front temperature sensor msg.device_id=1 ->msg.device_sn=‘machine1’ msg.device_id=2 ->msg.device_sn=‘machine1’ msg.device_id=3 ->msg.device_sn=‘machine1’ Data Source map group by device_sn latest status: speed,temp1,temp2 cache rear temperature sensor predictive model msg.defective_rate= func(speed,temp1,temp2) defective_rate>50%? trigger ,sms alert 1212
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.