Mainflux Labs is a technology company offering a unified IoT system consisting of integrated open-source IoT Cloud and Edge Computing technology, which can meet the requirements of the most complex IoT projects, solutions and products.

As the open-source IoT Platform, Mainflux is deployed in 14 countries around the world in various IoT solutions, including monitoring and controlling retail assets in Target (one of the biggest US retailers), Mauritius Telecom’s Towers, dwelling buildings in Berlin and Indian Coast Guard servers.

Mainflux Labs has implemented its open-source platform and developed custom components for numerous clients including, Worldwide provider of Oilfield & Gas Equipment – USA, System Integrator – Germany for pharma manufacturing machines, and EU Leader in Recycling and Waste Management Ventures.

Within SPATIAL, Mianflux participates in: WP – 1 Requirement and threat modeling; WP – 2 Resilient accountability metrics and embedded algorithmic accountability; WP – 3 System architecture, consistency and accountability for AI, Validation and Testing; WP – 4 User Engagement and Acceptance and WP – 6 Impact, Outreach and Collaboration.

Principal Investigator: MSc Ing. Ivan Milosevic

Ivan holds an MSc in Electrical Engineering from Belgrade University and is an IT expert with over 13 years of professional experience.

For ten years he worked in the three largest Serbian telecommunications companies (EUnet, SBB and mCloud) on a pioneering endeavour to introduce cloud hosting services in Serbia.

For the last 4 years he has been working at Mainflux Labs,  where he contributes to the development of the IoT platform using the most modern programming techniques, systems, libraries and frameworks – like Kubernetes, cybersecurity, IoT Edge and Cloud, AI/ML, WASM, distributed asynchronous systems and many others.

He has adapted and enabled the IoT platform for many clients who base their activities in various industries like public traffic, telecommunications and smart city.

In addition to professional involvement, he uses his experience and skills to help several civil society organisations in Serbia, which deal with environmental issues, participatory planning and access to common goods.

  1. How did you started/joined/enrolled SPATIAL?

Mainflux Labs has a long track record in IoT innovation. Mainflux IoT platform is an open-source, Apache-2.0 licensed software, hosted on the GitHub and open for collaboration to a wide community. Because of numerous contributors and it’s usage in many professional and public research projects, Mainflux platform became world-famous, and being regarded as one of the most sophisticated open-source platforms “made in Europe”.

All these are the reasons because of which Delft University invited Mainflux Labs in the consortium, when in SPATIAL project there was an evident need for IoT and 5G AI/ML security research. We accepted this invitation with pleasure and excitement.

  1. What are your expectations in a project of this nature?

Mainflux platform is open and it is growing with contributions of many researchers over the globe. There are a lot of features which we are trying to cover, and make it a new age of European smart cities and smart industry.

Security is one of the central points of Mainflux platform from the start. It is baked into the core design of the project. Following new standards and developments in this domain is an extremely important task. Additionally, Mainflux IoT platform is an infrastructure for sensor data acquisition and storage, but we are also guaranteeing security of this storage, as well as facilitation of the analytics and AI/ML training and inference that should be done on top. Finally, as we pitch Mainflux as an “Unified IoT Platform”, that works with the same codebase all over edge-fog-cloud continuum, it is important that AI algorithms are securely brought and deployed on the edge, using Mainflux as an infrastructure for orchestration and monitoring.

Through the SPATIAL project, we expect improvement in many areas of the product: cloud security, cloud-edge orchestration for AI workloads and specifically Confidential Computing aspects leveraging on Trusted Execution Environments.

  1. What can the research community expect from SPATIAL?

The research community will obtain results in in-depth research of AI security and explainability. Mechanisms for accountable and resilient AI using TEEs and novel Confidential Computing techniques will be one of the very interesting results. Evaluation testbeds will be very useful for the research community. Results from several use-cases will be described and will help better understand the usage of these novel techniques in practice. Finally, a set of documentation, tools and an educational module will help a research community to widen the knowledge in this area.

  1. Where do you see SPATIAL results in 10 years?

Privacy-preservation, accountability and resilience are extremely important for enabling wider use of AI, collaborative ML and data sharing for the purposes of training and inference. SPATIAL results, findings, tools, SDKs, documents, reports, metrics, Software techniques and other artefacts will help these improvements and AI acceleration in the decade to come.