Hello! I am helping several districts roll out Kolibri as an offline solution for students with zero internet access in rural Arkansas. Here are my thoughts and I would love some feedback/input as we look to roll this out.
A “primary” or “main” server would be housed on the school campus.
This one server would be the sole “Facility” in Kolibri
School teachers would have Coach accounts on the main server
School teachers would create “classes” - one class for each period or subject or however they see fit to organize their coursework. Most likely, teachers would have several classes they oversee.
ALL Coursework would be located on one server, in one facility
All students (roughly 200 initially, but up to 1100 possible) would be entered into the single facility.
Students would be assign to classes according to their schedules
Content would be propagated out to Raspberry Pi Zero Ws for each household (NOT each student).
Each student has a Chromebook or Tablet
Each student connects to their local/home-based Raspberry Pi device over the Kolibri wi-fi hotspot.
Students complete their work on their devices/local copy of Kolibri
Students return to campus at selected intervals to “sync” their data back to the main server on campus. (Hopefully implemented in 0.14, but can be done via command lines in 0.13 manually).
That is my plan at this point. One of the biggest issues I am told is the use of one single facility to handle all of this because of the potential for a vast amount of content being stored on all the memory cards for each Pi-ZeroW. Instead, it has been suggested that multiple Facilities be set up and associate students with a specific facility.
If we were to try to associate specific students with specific facilities, how much extra work is that in terms of making sure the right content goes to the right facility? I do not want to overload our teachers more than they are already being asked, if at all possible.
Thank you! It’s quite ambitious, and honestly pushes Kolibri outside of the scope of its design. I am aware of that, so the stumbling blocks we come across are expected (at least by me. We’ll see how the teachers feel). So far, the teachers a re generating content and curating content from the Studio channel providers.
I realize now that they are uploading their video lessons in far too high of a resolution, thus eating up valuable storage space plus wasting bandwidth. I will help them work through that.
In the meantime, I am working with the school’s local tech department to map out what the rollout looks like, what multiple facilities might look like, and rotating the Studio content against the local Kolibri content in order to not overload the Studio servers nor the local servers/student devices. There are a lot of plates starting to spin, for sure.
Since I cannot edit the original post, something that has changed is that we are looking at multiple facilities serving students in a targeted/specialized fashion. Not sure what that looks like as far as implementation just yet, but this will be better than trying propagate ALL the content to ALL the devices.
After looking at numbers of students needing OFFLINE content access, this project has been dramatically scaled back. Initially, we were expecting hundreds of students to have no online access. However, after evaluating several datasets, it appears we have far fewer students than anticipated.
The district has opted to go with an online solution for the majority of students. We will be rolling out Kolibri for those students without any access at all.
It is my hope that one day, I help oversee a large rollout as the one we thought we would be doing here. In the meantime, my very sincere thanks to the Kolibri team and the community for helping me work through some of the challenges as we move forward with this revamped approach to using Kolibri in a way more in line with its original intent.
