Liqo 0.1 is out in the wild!

In this first release, we realized a proof of concept of the experience Liqo would like to provide to users. In particular:

• Dynamic discovery and peering between clusters: clusters can be discovered and peered.
• Transparent pod offloading in a multiple cluster scenario with no modifications required w.r.t. applications and Kubernetes Control Plane.
• Dynamic and decentralized network extension across clusters, supporting common traffic patterns (Pod to Pod, Pod to Service).

Take a look at the source code on GitHub.

The idea for Liqo grew from a struggle we had at our University

We had plenty of powerful computers in our labs, but there were always some students complaining about their laptops not being powerful enough to run their jobs.

Our idea was straightforward: let’s transform our labs, including student laptops, into opportunistic data centers, so that heavy tasks can always be executed on the best resource among those available.

However, students didn’t want to give up the control of their laptops nor compromise in security. Here is where Liqo comes in: a way to share resources and services, while each user retains the full control on his/her box, whether it is the lab IT admin or the student, deciding what to share, how much, with whom. Controlled, secure sharing.

While working on this project, we realized that the capability to share Kubernetes clusters has much greater implications. Liqo can not only solve our student’s problems but also become the enabler for applications that rely on the democratization of edge computing such as Autonomous driving, IoT, Industry 4.0, Smart cities.

In fact, Liqo allows to run services on multiple clusters without even having to provision, build and operate your own distributed edge infrastructure. This project could optimize utilization of currently available resources and kickstart a first scale up of edge computing.