Kenobi
Intelligent SD-WAN platform born from a collaborative R&D (Innoviris + VLAIO)
Client
- KERN-IT x VENN Telecom
- Telecom & R&D
- 2020 -2022
Services
- BackEnd Development
- Data Science
- FrontEnd Development
- GIS
- IoT / Edge Computing
- SDWAN
VENN Telecom is a Belgian operator of hybrid enterprise connectivity (SD-WAN, 4G/5G, satellite), specialised in networking hard-to-connect sites. In 2020, a shared observation with KERN-IT: managing ever larger, distributed and heterogeneous SD-WAN, IoT and Edge fleets had become impossible with fragmented, purely reactive vendor tools.
KERN-IT and VENN then launched Kenobi, a 24-month research project co-funded by Innoviris and VLAIO (BEL SME / BEL-COO programme), to prototype an intelligent, vendor-agnostic network management platform. Unlike vendor tools centred on a single ecosystem, Kenobi aims for a cross-cutting software layer, business-oriented and independent of brands.
The project supervised more than 2,000 network points across Europe in real conditions, and gave birth to bricks now used in production, including the GISWAN network mapping tool. Beyond the deliverables, Kenobi grew KERN-IT's expertise in telecom, data and edge, and proved its ability to turn a demanding subsidised R&D into concrete value.
A node management platform
The foundation of the project: a platform able to absorb large volumes of heterogeneous data and integrate, in real time, thousands of devices spread across Europe through dedicated connectors.
It centralises what used to be scattered:
- Supervision of a multi-site, multi-operator SD-WAN fleet
- Real-time import and update of thousands of nodes
- A cross-cutting view, independent of manufacturers
A base designed from the start to scale to several thousand units.
Automation and intelligence
At the heart of the research: turning network data into actions. The teams prototyped a flexible alerting system, a machine learning model for problem detection, and an action matrix for automated incident resolution.
The work focused on:
- Network problem detection through machine learning
- Exploration of self-healing mechanisms
- A conversational bot for device management
With a key lesson: understanding where automation adds value, and where humans remain essential.
Network, bandwidth and cost optimisation
To give an enriched view of a device's state, the platform analyses key network metrics in real time, collected through the nodes and operator APIs.
The algorithms developed enable:
- Real-time signal analysis: RSSI, packet loss, 3G/4G/5G
- Optimisation of link quality and stability
- Choosing the best available antenna for each connected object
The goal: connectivity that is more reliable and less costly, without constant human intervention.
Geographic data, the birth of GISWAN
One of the project's most structuring results. From geographic and environmental data, KERN-IT developed GISWAN, a tool to visualise network quality in the field.
GISWAN makes it possible to:
- Visualise the quality of signals collected in the field
- Overlay antennas, coverage and observed performance
- Support decisions before and during complex rollouts
First a research brick, GISWAN is now a development axis in its own right, presented to industrial prospects.
Edge Computing and 5G
The project's final frontier: mastering the boundary between cloud and field. The teams designed an edge computer connected to SD-WAN equipment and experimented with managing it remotely.
The experiments covered:
- Remote deployment and update of nodes
- Containerisation and orchestration at the edge
- Deciding where intelligence belongs, cloud or edge (latency, bandwidth, reliability)
Strengthening KERN-IT's ability to design pragmatic, robust Edge architectures.
