Intelligent Magic Application
If you don't make a product that's good for developers then the developers won't make a good product.
In a Nutshell
- Allocates a VM with a minimal Ubuntu installation and configures it according to specification.
- Orchestrates a Kubernetes cluster with modifications that's needed to run PCC.
- Builds the mobility images and PCC helm charts in the designers repo.
- Deploys PCC with flags suitable for the cluster.
- Configures Dallas and Mobility to match each other to be able to run traffic.
IMA was developed at the very early experiments on getting the MME to run on Kubernetes.
Primarily back then the focus was on experimenting with the configuration of Kubernetes and the host to be able to run on it.
- Built by developers for developers.
- Pre-commit feedback loop.
- One button from nothing to traffic-ready.
- Contain executable documentation on how to configure vanilla Linux and Kubernetes to be able to deploy PCC.
The core of IMA functionallity is the ability to write scripts that executes code on different machines.
And to try to be as transparent as possible both in the source code and in the logging.
This is how the typical IMA instance is allocated:
In this table I try to describe other tools that exists in this space and try to map out what they are in charge of deciding.
|What machine to run on||X||X||X|
|Making PCC build||X||X||X|
IMA source code is quite mature, not growing rapidly.
- Aleksander Zhitnyakov
- Aleksandra Kulakova
- Alexander Wingård
- Colin Taylor
- George Zachariadis
- Hanbing Zhang
- Henrik Johansson
- Johan Sandstig
- Johan Svahn
- Juan Ricart
- Karl Eberth
- Karol Ostrovsky
- Kirill Malinin
- Marco Zhang
- Martin Tengklint
- Mats Jansborg
- Miguel Gargallo
- Oliver Kjellman
- Patrik Stoppenbach
- Per Bohlin
- Pierre Krafft
- Sveinn Finnsson
- Syafiq Atiiq
- Urban Boquist
- Yohanes Kuma