Boundless stack

Overview

Boundless Geospatial provides commercial support for open source geospatial software.

They put together an open source stack and then commercialized it. Then Planet Labs bought them.

Back a couple years ago I talked to them at a FOSS4G conference and they seemed so bright-eyed and helpful. I am sad today because every time I find material about them I hit a paywall when I try to follow up on it.

This is the business model that works best for me, a passionate after-hours GIS tinkerer:

1. A free tier. Unsupported. Developers get access and I can stand up a server on my own with low traffic levels at no cost.
2. A mid-level tier for my clients, a reasonable $ subscription price and limited support.
3. An enterprise tier that costs $$$ with full support. I never work in this tier so the cost matters naught to me.

My impression is Boundless pretty much built up the GeoServer I have come to know and love.

I guess at this point I just ignore them until I hear something has changed.

Github: The sea of lost projects

Github is a Sargasso Sea of projects and Esri and Boundless have thrown many into the morass. I have too but I am just some random guy using their servers as place to tether my code so I can't misplace it. I am not claiming to be an Enterprise Leader. Anyway.

I tried a few boundless/* things until I was ready to give up. Things like boundless/composer with 100's of security faults.

What is in the Boundless stack?

1. QGIS on the desktop
2. Boundless Composer is for authoring maps
3. OpenLayers is a JavaScript library for building web maps.
4. GeoServer http://geoserver.org/ Front end for geospatial repository https://github.com/Geo-CEG/docker-geoserver
5. GeoGig http://geogig.org/ There is a GeoServer plugin for change management https://github.com/Geo-CEG/docker-geogig
6. GeoWebCache http://geowebcache.org/ will cache pre-rendered tiles coming from Geoerver
7. PostGIS Used as a data repository for GeoServer but capable of full SQL processing of spatial data.

See this page: Boundless committed to open source and their github page: https://github.com/boundlessgeo (where you will find they have committed code but failed to maintain it.)

I would consider adding OpenDataKit to this stack.

Maybe VirtualBox will work

...or maybe when I look for it I will hit a paywall. :-(

How building GeoNode fails

Download and build

cd source/docker
git clone [email protected]:boundlessgeo/geonode.git
cd geonode
docker-compose build

This build includes

• postgres
• elasticsearch (had to modify the compose file as there is no elasticsearch:latest tag, I used 6.6.0)
• rabbitmq
• django
• celery
• geoserver (built on tomcat 9)
• geonode (built on geonode/nginx)
• ArcREST

It keeps its data in a data container called geoserver_data_dir

docker-compose up
.
.
.
Creating geonode_rabbitmq_1      ... done
Creating geonode_postgres_1      ... done
Creating geonode_elasticsearch_1 ... done
Creating geoserver_data_dir      ... done
Creating geonode_geoserver_1     ... done
Creating geonode_celery_1        ... error
Creating geonode_django_1        ... error

Who needs django and celery anyway? Oh right -- that's basically the underpinnings what I wanted to test today, geonode. That makes Boundless currently a non-starter. Trying GeoNode official version instead!

There's a quote regarding open source about "the smartest people don't work in your company" and I think that's true about Boundless. OKAY to be fair, I could spend perhaps another day and get it going. But I just don't know if I even want to use GeoNode yet. I probably don't.

The thing is, if I contracted with Boundless doubtless I could spend 4 hours on the phone and they'd help me set it up, but first I'd have to be convinced I was not just substituting Boundless for ESRI and just beholden to a new master.