My current project at Heata is coming into a live Trail and I thought it would be fun to share how we are using LXD. A huge thank you to everybody that helped me with my learning curve and problem solving developing this solution it has made my development that much easier.
At its core its using LXD containers to provide customer sandboxed compute applications on a distributed network of compute nodes inside domestic homes.
The concept is a climate friendly distributed data centre where hat generated by the nodes is using by the domestic home user. Thus this technology can offset the energy used by compute nodes with the energy normal used to heat domestic hot water. As currently CO2 from data centres exceeds 3% of the total global emissions this project is a chance to move away from large energy hungry data centres which are due to demand are due to double in size in the next 5-10 years.
The challenges with this architecture are physical security so we have LUKS2 encrypted drives, Secure boot and TPM2 so that the applications and sensitive data on the appliances are safe.
Secondly because this phase of the solution currently works over the homes domestic broadband they are engineered to work standalone with occasional and low bandwidth polls to a controller serve which gives them jobs. PyLXD was particularly useful for implementing a task/job management for orchestration.
We are currently targeting the solution for inclusion for local council made social housing which would supply the occupant with free hot water and broadband.
Any questions please fire away the glossy information for this solution is below:-
Technical information here:-