In-network computing for Industry 4.0 and beyond
Fast is never fast enough. Especially not in the Industrial Internet of Things (I-IoT). The process for sensors to detect the environment transmitting a vast volume of data towards the cloud to analyse them and act upon it, takes time. Now, there is a possibility to reduce this latency. EIT Digital has an industrial doctorate position available at Ericsson Hungary and ELTE University in Hungary, in where you can work on a new paradigm called in-network computing. Interested to compute in the network? Then apply now.
Sub-millisecond latency is utmost important in low-level real-time industrial control of robots and sensor data-based event processing in industrial environments. Currently, programmable Logic Controllers (PLC’s) are used for real-time industrial control of robots. Performing sensor data processing, filtering and event processing can however be much faster than the currently applied server-based solutions.
Therefore, the industrial doctorate candidate will develop novel methods demonstrating how P4-enabled networking devices, like switches, smart NIC’s and distributed service cards, can be used. (S)he will show in practice that these programmable networking devices provide flexible alternatives to rigid PLCs.
Since the devices can be programmed in the data plane programming language called P4, new features can be added easily, opening the door to faster innovation in the lower layers of industrial control.
It is expected that moving the low level intelligence of industrial compliances, including robot arms, mobile robots, and conveyor belts, to the network like a 5G Edge Computing node with a P4-enabled component, can solve general issues like precise timing and synchronization of various industrial processes, ultra-fast reactions to emergency situations. It can also enable the plug-and-play operation of smart factories since the control from low to high levels are implemented outside of the device.
In short, the industrial doctorate candidate will face three challenges to solve:
- time sensitive in-network computation,
- reliable in-networking computing, and
- energy efficiency, fault tolerance, security and trust aspects of in-networking computing.
Computing in the network is a new research area that has emerged with the advent of programmable switching ASICs over the past few years. In-network computing refers to the execution of programs typically running on end-hosts within network devices. The computation in this model is done within the network, using devices that have already been deployed and are already used to forward network traffic. In contrast to traditional network computing where the computation is done by computers deployed in a network, the computations are performed by programmable switches. The two key benefits of in-network computing are high throughput and ultra-low latency.
The key innovation potential in this PhD is that programmable P4-enabled networking hardware can replace the inflexible PLC-based systems in nowadays industrial setups. The solution to be developed in the thesis work will be an evolutionary step towards a fully softwarised, more flexible industrial automation, solving both low and high-level control tasks in software.
In-networking computing is a new field in the telecommunication and industrial automation landscape. Therefore, it is expected that the PhD work will lead to 1-2 patents every year. Other innovations to expect are the implementation of the methods in working prototypes based on P4-enabled hardware (Barefoot Tofino, smartNICs, DSCs) and integration of them into Ericsson Flight Rack product. Also, the interfacing with existing industrial controllers like ROS-I, Ericsson’s Industry Connect will be solved. Furthermore, the EIT Digital industrial doctorate will work on early stage use cases for both the short and the long term. (S)he will give live demonstrations of the prototypes at several events.
The expected academic outcomes the PhD candidate are:
- 2 top-level conference publications
- 2 high-impact journal submissions
- patent applications
The industrial doctoral student involved in this programme will be work from his base location at the Co-Location Centre of the EIT Digital Budapest Node. Furthermore (s)he will share her or his time at the premises of Ericsson Hungary and ELTE university. As part of the mobility plan of the industrial doctorate the student might have the chance to work at KTH Royal Institute of Technology in, Sweden, the Technical University of Munich (TUM) in Germany, the Oxford University in the UK and/or the ETH Zürich, Switzerland. In addition to the industrial doctorate research, the PhD student will also be following leadership seminars.
Those interested in applying should complete the EIT Digital Doctoral School application form.
Please apply before 9 October 2020.