Hi Losant team,
I think it would be really great if you were able to provide a new section in the LHS panel for creating and deleting Docker instances of your Edge GW (agent) as new devices with some level of CLI access for managing these Docker images, and exporting them as well for use on physical HW if required. Using the more open MQTT broker on the Edge GW is very convenient testing topics that are prohibited in the core losant broker, and there are a number of input types like UDP & TCP that only the Edge agent/GW can support, and being able to test these from within the Losant dev environment in its entirety would be really useful. You’d likely have to allow assignment of a publicly reachable static IP address for IP ingress connectivity, and an admin feature to request and manage these (i.e .FW, add/delete), but it would make developing and testing Edge Workflows a whole lot simpler before any HW needed to be prepped.
This is an interesting idea, and I’d love to learn more about your use case. Are you looking to, from the cloud, spin up an Edge Agent instance on a specified gateway?
but it would make developing and testing Edge Workflows a whole lot simpler before any HW needed to be prepped.
The Edge Agent can run anywhere that Docker can run. Personally, I have an Edge Agent instance running on my laptop that I use for testing purposes. Is this something that you’re able to do? Is there something with this process that is not providing what you’re looking for?
While we have talked about adding a cloud-based environment for testing Edge workflows, we’ve also questioned just how useful it would ultimately be since the configuration for all the southbound connectors is very hardware and network-dependent, and that’s not something we can simulate for users.
Your post has stirred up some internal conversation, so I will be sure to make this feature request for you and let you know what I learn.
Hi Heath,
the problem with spinning up a Docker edge instance on external HW is all the effort needed to go to around networking and firewall config etc. For example, if I run a Docker Edge agent on my home server, and need to test a cellular mobile data device connecting to it using MQTT, UDP, or TCP, I have to figure out what the mobile operator’s IP ranges are, if any NAT is being used, and set up DNS resolution for the Edge agent target hostname to point to my (usually variable/transient) home broadband provided ingress IP address, and also figure out the inbound FW config needed. In my case, the mobile operator has 4 different IP subnet pool ranges that could be assigned to the cellular IoT device, so that makes for a complex FW ruleset. It’s even worse in corporate NW scenarios, with much stricter NW & FW ingress change control procedures. If I try to avoid all that, my next option is to set up a VM or Docker container in Kubernetes in Google GCP, Azure, or AWS, and that’s an awful lot of config work for a simple need to live-test a cellular IoT device with an Edge workflow, before moving on to building the Losant Experience app.
So if Losant had a safe per-Org Docker runtime option available to bring up an Edge Agent, I could avoid all of that and get on with my core task of building and testing the desired Edge workflows.