EMQ X Kuiper is an edge lightweight IoT data analytics / streaming software implemented by Golang, and it can be run at all kinds of resource constrained edge devices. One goal of Kuiper is to migrate the cloud streaming software frameworks (such as Apache Spark,Apache Storm and Apache Flink) to edge side. Kuiper references these cloud streaming frameworks, and also considered special requirement of edge analytics, and introduced rule engine, which is based on Source
, SQL (business logic)
and Sink
, rule engine is used for developing streaming applications at edge side.
User scenarios
It can be run at various IoT edge use scenarios, such as real-time processing of production line data in the IIoT; Gateway of Connected Vehicle analyze the data from data-bus in real time; Real-time analysis of urban facility data in smart city scenarios. Kuiper processing at the edge can reduce system response latency, save network bandwidth and storage costs, and improve system security.
-
Lightweight
- Core server package is only about 4.5M, initial memory footprint is about 10MB
-
Cross-platform
- CPU Arch:X86 AMD * 32, X86 AMD * 64; ARM * 32, ARM * 64; PPC
- The popular Linux distributions, OpenWrt Linux, MacOS and Docker
- Industrial PC, Raspberry Pi, industrial gateway, home gateway, MEC edge cloud server
-
Data analysis support
- Support data extract, transform and filter through SQL
- Data order, group, aggregation and join
- 60+ functions, includes mathematical, string, aggregate and hash etc
- 4 time windows
-
Highly extensibile
Plugin system is provided, and it supports to extend at
Source
,SQL functions
andSink
.- Source: embedded support for MQTT, and provide extension points for sources
- Sink: embedded support for MQTT and HTTP, and provide extension points for sinks
- UDF functions: embedded support for 60+ functions, and provide extension points for SQL functions
-
Management
-
Integration with EMQ X Edge
Seamless integration with EMQ X Edge, and provided an end to end solution from messaging to analytics.
-
Pull a Kuiper Docker image from
https://hub.docker.com/r/emqx/kuiper/tags
. -
Set Kuiper source to an MQTT server. This sample uses server locating at
tcp://broker.emqx.io:1883
.broker.emqx.io
is a public MQTT test server hosted by EMQ.docker run -d --name kuiper -e MQTT_BROKER_ADDRESS=tcp://broker.emqx.io:1883 emqx/kuiper:$tag
-
Create a stream - the stream is your stream data schema, similar to table definition in database. Let's say the temperature & humidity data are sent to
broker.emqx.io
, and those data will be processed in your LOCAL RUN Kuiper docker instance. Below steps will create a stream nameddemo
, and data are sent todevices/device_001/messages
topic, whiledevice_001
could be other devices, such asdevice_002
, all of those data will be subscribed and handled bydemo
stream.-- In host # docker exec -it kuiper /bin/sh -- In docker instance # bin/cli create stream demo '(temperature float, humidity bigint) WITH (FORMAT="JSON", DATASOURCE="devices/+/messages")' Connecting to 127.0.0.1:20498... Stream demo is created. # bin/cli query Connecting to 127.0.0.1:20498... kuiper > select * from demo where temperature > 30; Query was submit successfully.
-
Publish sensor data to topic
devices/device_001/messages
of servertcp://broker.emqx.io:1883
with any MQTT client tools. Below sample usesmosquitto_pub
.# mosquitto_pub -h broker.emqx.io -m '{"temperature": 40, "humidity" : 20}' -t devices/device_001/messages
-
If everything goes well, you can see the message is print on docker
bin/cli query
window. Please try to publish another message withtemperature
less than 30, and it will be filtered by WHERE condition of the SQL.kuiper > select * from demo WHERE temperature > 30; [{"temperature": 40, "humidity" : 20}]
If having any problems, please take a look at
log/stream.log
. -
To stop the test, just press
ctrl + c
inbin/cli query
command console, or inputexit
and press enter. -
Next for exploring more powerful features of EMQ X Kuiper? Refer to below for how to apply EMQ X Kuiper in edge and integrate with AWS / Azure IoT cloud.
- Using JMeter MQTT plugin to send simulation data to EMQ X Broker, such as:
{"temperature": 10, "humidity" : 90}
, the value of temperature and humidity are random integer between 0 - 100. - Kuiper subscribe from EMQ X Broker, and analyze data with SQL:
SELECT * FROM demo WHERE temperature > 50
- The analysis result are wrote to local file by using file sink plugin.
Devices | Message # per second | CPU usage | Memory usage |
---|---|---|---|
Raspberry Pi 3B+ | 12k | sys+user: 70% | 20M |
AWS t2.micro( 1 Core * 1 GB) Ubuntu18.04 |
10k | sys+user: 25% | 20M |
- 8000 rules with 800 message/second
- Configurations
- 2 core * 4GB memory in AWS
- Ubuntu
- Resource usage
- Memory: 89% ~ 72%
- CPU: 25%
- 400KB - 500KB / rule
- Rule
- Source: MQTT
- SQL: SELECT temperature FROM source WHERE temperature > 20 (90% data are filtered)
- Sink: Log
- Go version >= 1.11
-
Binary:
-
Binary:
$ make
-
Binary files that support EdgeX:
$ make build_with_edgex
-
-
Packages:
$ make pkg
-
Packages:
$ make pkg
-
Packages files that support EdgeX:
$ make pkg_with_edgex
-
-
Docker images:
$ make docker
Docker images support EdgeX by default
To using cross-compilation, refer to this doc.