IoTan is a set of components and infrastructure for IoT data acquisition, storage and analysis.
IoTan core is developed is scala and can leverage existing Java libraries that help building domain components, such as haystack-java.
In version 0.1, IoTan uses Haystack protocol and concepts as the primary way to ingest and serve data.
Clients can interface to IoTan using haystack client or scala based API from scala ir sbt console.
IoTan Core currently uses Cassandra database as store (a future version may support MongoDB). Before IoTan can be made operational, Cassandra has to be setup.
Note: some iotan-related names within the project use "iotus" instead of "iotan", which is an older name of the project. This will eventually change.
In version 0.1, IoTan offers Cassandra as data store. It can be deployed on a single virtual or physical machine in development mode, on two machines on Docker conainters, or on several machines in production, in which case Cassandra cluster of 3 or more nodes would be used.
Dev Deployment Example:
Production Deployment Example:
Please see roadmap below for future architecture and feature plans.
See http://planetcassandra.org/cassandra/
Note on cqlsh:
With some version of cqlsh, when connecting to C* 3.9, you need to specify cqlversion option:
cqlsh --cqlversion="3.4.2"
cqlsh
CREATE KEYSPACE IF NOT EXISTS test WITH replication = { 'class': 'SimpleStrategy', 'replication_factor': 1 };
For Cassandra test to pass, a test and iotus keyspaces needs to be created and the databased initialized. For LDAP tests to succeed, LDAP server has to be setup.
-
For Cassandra IoTan database, create iotus keyspace and init the db:
cqlsh [hostname] CREATE KEYSPACE IF NOT EXISTS iotus WITH replication = { 'class': 'SimpleStrategy', 'replication_factor': 1 };
-
Then run schema and populate script. See ./db
cqlsh [hostname] < db/iotus-schema.cql cqlsh [hostname] < db/iotus-populate.cql
You can execute cqlsh commands like in this example (using hostname u16 and keyspace iotus)
# set proper path to CQLSH if needed
#export CQLSH=~/.ccm/repository/3.9/bin/cqlsh.py
export CQLSH=cqlsh
# create namespace for testing with one host (for production, see
# iotus-keyspace-prod.cql
# $CQLSH u16 -k iotus_simple -e "USE iotus_simple;"
$CQLSH u16 -e "CREATE KEYSPACE iotus WITH replication = { 'class': 'SimpleStrategy', 'replication_factor': 1 };"
for f in iotus-schema.cql iotus-populate.cql; do
$CQLSH u16 -k iotus_simple -f $f
done
-
Build library jar
sbt package
The resulting files will be placed in target, e.g. like this: ./target/scala-2.11/iotus-core_2.11-0.1.01-SNAPSHOT.jar
sbt compile: compile sources
sbt test: run tests
sbt console: prepare for command line scala execution
-
package all dependency jars. This can be used to execute from a java project, such as iotan-hs Resulting jars will be in ./target/pack/lib
sbt pack
ls -l ./target/pack/lib
total 132248 -rw-r--r-- 1 peter staff 2064728 Apr 21 2014 argonaut_2.11-6.0.4.jar -rw-r--r-- 1 peter staff 990392 Jan 28 2016 cassandra-driver-core-3.0.0.jar -rw-r--r-- 1 peter staff 575389 Apr 14 2008 commons-collections-3.2.1.jar -rw-r--r-- 1 peter staff 354491 Jan 27 2012 commons-configuration-1.8.jar ...
-
assembly: package all dependencies into one fat-jar This needs work since it generates an exception related to dupliates in io.netty:
sbt assembly ... java.lang.RuntimeException: deduplicate: different file contents found in the following: /Users/peter/.ivy2/cache/io.netty/netty-handler/jars/netty-handler-4.0.33.Final.jar:META-INF/io.netty.versions.properties /Users/peter/.ivy2/cache/io.netty/netty-buffer/jars/netty-buffer-4.0.33.Final.jar:META-INF/io.netty.versions.properties ...
If all built and db has been created and populated like above, then make sure iotus.yaml config file is setup properly, e.g. pointing to the proper db url, something like:
db: cassandra://localhost:9042/iotus
or
db: cassandra://localhost:u16/iotus_simple
and then run from sbt console:
import iotus.core.IMContext
import iotus.core.IMFrame
// create a context and perform a query
val ctx = new IMContext("test-project", "[email protected]", "point").readAll()
ctx.toGrid.dump
// try another query
ctx.clear.readAll("site").toGrid.dump
If database errors are displayed, it means the db has not been properly initialized.
Note: this is needed only if you prefer LDAP autentications instead of database user table based authentication.
Use OpenLDAP related resource to install LDAP.
After installation, verified connectivity before setting up parameters in your config file iotus.conf and test file src/resources/application.conf
E.g. use ldapsearch to verify. Here is an example (replace 'admin', example, uid etc. with correct values:
ldapsearch -w {adminPassword} -h ldap.example.com -D 'cn=admin,dc=example,dc=com' -b 'dc=example,dc=com' "uid=frank"When deploying for the first time on development or production system, create iotan user, to run as instead of root, a target configuration and related directory structure.
As sudo root, create iotan user and setup configuration
adduser iotan
# folow the instructions to setup use account
mkdir -p /opt/iotus
chown iotan:iotus /opt/iotus
# as user iotus
mkdir -p /opt/iotus/conf
cp conf/iotus-template.yaml /opt/iotus/conf/iotus.yaml
# adjust configuration to point to the correct db and possibly ldap
vi /opt/iotus/conf/iotus.yaml
There will be two docker container needed, one for cassandra, using standard docker library and one for IoTan core and haystack server.
Therefore the docker runtime container deployment may look like this, with two docker containers on one host.
Docker Dev Deployment:
Also see
./docker
./lucene
To run a single development instance of Cassandra via docker:
# --- start cassandra container and copy lucene jar
# If this is the first time on this docker system
# it will take a while to download
docker run --name cass01 -d cassandra:3.9
# copy lucene jar file for lucene text search indexing support
docker cp lucene/cassandra-lucene-index-plugin-3.9.5.jar cass01:/usr/share/cassandra/lib/
# restart container
docker restart cass01
# --- create keyspace and schema
# a) first, copy files to docker container
# use docker container name from above, such as cass01
CASS_NAME=cass01
docker cp docker/populate_db.py $CASS_NAME:/tmp
docker exec -i $CASS_NAME sh -c 'mkdir -p /tmp/dev'
for f in docker/dev/*; do
docker cp $f $CASS_NAME:/tmp/dev
done
# b) execute populate
docker exec -i $CASS_NAME sh -c '/tmp/populate_db.py -s localhost iotus'Alternatively, the run command above could look like this if you'd like to preserve and archive the data on the host:
docker run --name cass01 -v /my/own/datadir:/var/lib/cassandra -d cassandra:3.9
To run IoTan core and Haystack server from docker repo.
docker run --link cass01 --env CASSANDRA_HOST=cass01 -p 1225:1225 --name iotan01 -d petermic/iotan:0.1
Default values for environment variables: CASSANDRA_HOST=cassandra
The docker run command above will start iotan-core service and iotus haystack server, connecting to the running cassandra container cass01. During initialization, it will create proper /opt/iotus/conf/iotus.yaml file for that and will initialize cassandra keyspace "simple" and pre-populate it will a default user, admin id and password as specified and project "simple" as described in the documentation.
With the arguments above, the haystack server will be available on the docker IP for that instance at port 1225 and can be tested like this. Note the default password for the user [email protected] setup by docker auto-setup. Another user is also setup (see below)
# find out the IP
SIPD=$(docker inspect --format {{.NetworkSettings.IPAddress}} iotuscore01)
# perform test via un-authenticated about request
curl http://$SIPD:1225/api/about
# it can do an authenticated request to query points, like this:
curl -u [email protected]:4dmin123 http://$SIPD:1225/api/simple/read?filter=equip
# you will get a response tht looks something like this:
ver:"2.0"
id,equip,tz,navName,siteRef,origId,dis
@2c34a5f0-e329-4f6e-a679-00ac3b97ee3a "Office RTU-1",M,"Las_Vegas","RTU-O-01",@a688156f-9584-49ec-9fb2-dc3b3e565bd6,"Office.RTU-1","Office RTU-1"
@9d244671-4216-4578-97e6-208d04a96ab9 "Office RTU-2",M,"Las_Vegas","RTU-O-02",@a688156f-9584-49ec-9fb2-dc3b3e565bd6,"Office.RTU-2","Office RTU-2"
@cb24697d-0f7d-403c-bd05-5a9ae05d53cf "Home ElecMeter-Main",M,"Los_Angeles","ElecMeter-Main",@32cb5eca-b8b1-47cd-88ee-b9fd26e1f4b6,"Home.ElecMeter-Main","Home ElecMeter-Main"
# or
curl -u [email protected]:4dmin123 http://localhost:1225/api/simple/read?filter=site
# you will get a response tht looks something like this:
ver:"2.0"
id,area,geoAddr,geoCity,site,tz,geoState,geoStreet,geoPostalCode,origId,dis,geoCountry
@32cb5eca-b8b1-47cd-88ee-b9fd26e1f4b6 "Home",3000,"10800 Torre Ave, Cupertino, CA 95014","Cupertino",M,"Los_Angeles","CA","600 W Main St","95014","Home","Home","US"
@a688156f-9584-49ec-9fb2-dc3b3e565bd6 "Office",6000,"900 N Roop St, Carson City, NV 89701","Carson City",M,"Las_Vegas","NV","3504 W Cary St","89701","Office","Office","US"
When iotuscore docker container is started for the first time via "docker run" command, the entrypoint.sh is invoked which peforms three actions:
- initializes the database if needed
- populates a sample project called "simple" if it doesn't exist yet
- start haystack server
The first two steps may be skipped it it's detected the database already contains the needed information.
Initializing the database consists of these steps:
- keyspace is created if it doesn't exist yet
- schema if created for the given keyspace
- two users are created: admin and guest ([email protected], [email protected])
For the sample project "simple", guest has read access and admin has read/write access (i.e. haystack hisWrite will work only for the admin user).
The "simple" project has several site, equip and point nodes that can be queried via iotus scala API or haystack client
If needed, additional project with the same content could be created via ProjectBuilder script, like this (this uses "myproject" project name and an existing sample zinc file to populate a few site, equip and point records:
./srepl.sh docker/project_populate.scala myproject America/Log_Angeles ./src/main/resources/iotus-simple.zinc false
To build the docker image for IoTan core and haystack server, you'll first need to create supporting tarballs, like this:
-
build core and hs projects and dependencies, e.g. like this:
cd iotan-core sbt compile pack ./dpack.sh cd ../iotan-hs ./gradlew build && ./deploy.sh ./dpack.sh cd ..
-
copy binaries to docker directory
cp iotan-core/iotus-core-bin-0.1.01.tar.gz iotan-core/docker cp iotan-hs/iotus-hs-bin-0.1.01.tar.gz iotan-core/docker cp iotan-hs/iotus-hs-bin-deps-0.1.01.tar.gz iotan-core/docker
-
download scala 2.11.8 from https://www.scala-lang.org/download/2.11.8.html to docker directory
ls -l -rw-r--r-- 1 peter dev 28678231 Mar 4 2016 scala-2.11.8.tgz
-
now build the docker, e.g. like this
cd iotan-core/docker docker build --tag myiotan:0.1 .
docker run -it --link cass01 --env CASSANDRA_HOST=172.17.0.2 --rm petermic/iotan:0.1 sh -c 'exec /opt/iotus/iotan-hs/srepl.sh'
import iotus.core.IMContext
val ctx = new IMContext("simple", "[email protected]")
#val ctx = new IMContext("demo", "[email protected]")
ctx.readAll("point").toGrid.dump
# you'll get a frame back that looks something like this:
ver:"2.0"
id,equipRef,point,his,tz,dis,kind
@756716f7-2e54-4715-9f01-91dcbea6c111 "demo pt 01",@756716f7-2e54-4715-9f01-91dcbea6c555,M,M,"New_York","demo pt 01","Number"
@756716f7-2e54-4715-9f01-91dcbea6c222 "demo pt 02",@756716f7-2e54-4715-9f01-91dcbea6c555,M,M,"New_York","demo pt 02","Number"
@756716f7-2e54-4715-9f01-91dcbea6c223 "demo pt 03",@756716f7-2e54-4715-9f01-91dcbea6c555,M,M,"New_York","demo pt 03","Number"
IoTus uses Project Haystack concepts and java haystack open source server and library for ingestion and output via haystack channel.
After version 0.1, the following features are planned for v 0.2 and beyond:
- simple CLI commands to setup and manage projects, and import .csv or .zinc data
- fine-grained group based permissions for project access
- tools for data scientists to analyze data
- add-on services to ingest and process data
- based on feedback possible support for MongoDB, in addition to Cassandra as a more lightweight deployment
Apache 2.0