Communication Networks Laboratory (CNL) Research Group

Description of Activities

ELTE CNL was originally established by ELTE and Ericsson Research. Since the beginnings the research group’s scope has shifted towards network programmability, in-network computing and quality of service assurance. The core activities are related to 1) data plane programmability for ensuring deep slicing architecture for 6G, 2) programmable QoS framework for SLA enforcement and measurements in cellular networks and data centers, 3) in-network computing for traffic reduction and sensor stream processing in industrial and other scenarios, 4) in-network robot control, 5) P4 code generator for offloading computing tasks to the network, 6) explainable graph neural networks for root cause analysis in cellular networks, 7) network programmability using the big switch model, 8) P4 program aggregation and disaggregation, 9) 3D point cloud processing with low complexity feature descriptors. Some of the activities are carried out in close cooperation with Ericsson Research, but we also have joint activities with universities from EU and US.

Research Interests

Programmable data planes, in-network computing, QoS and resource sharing, network programmability and automation, reinforcement learning and graph neural networks for networking applications.

Research infrastructure

Our testbed facility: 1x Stordis BF2556-1T (2 TBPS Tofino-1 P4 programmable switch), 1x APS6232D (6.4 TBPS, Tofino-2 P4 programmable switch – under procurement), 2x Intel Xeon server (dual socket Intel Xeon Silver 4110, 2.10GHz, 128GB RAM, 2x10GbE Intel NICs), 2x AMD server (AMD Ryzen Threadripper 1900X 8C, 128GB RAM, 2x10GbE Intel NICs, 2x100GbE Chelsio NICs) and 1x AMD server with GPUDirectRDMA (AMD Ryzen Threadripper 3970X 32C, 64GB RAM, 2x100GbE Mellanox ConnectX-5, NVidia Quadro GPU). We also have access to other testbed facilities located at our academic and industrial partners. 

Research Staff

• Sándor Laki, head of CNL (web: https://lakis.web.elte.hu, email: lakis@inf.elte.hu) 
• Gergő Gombos, senior member 
• Péter Vörös, senior member 
• János Szalai-Gindl, senior member 
• Dániel Varga, PostDoc researcher 
• Csaba Györgyi, PhD student 
• Dávid Kis, PhD student 
• Hiba Mallouhi, PhD student 
• Károly Kecskeméti, MSc student 

Projects

• HORIZON-JU-SNS-2023 NATWORK (300.000 EUR)
• Horizon Europe 6GSNS DESIRE6G (2023-2025, 365.000 EUR) 
• OTKA FK Foundations of data plane program analysis (2021-2025, 120.000 EUR) 
• H2020 NGI Pointer OC2 - P4EDGE (2021-2022, 200.000 EUR) 
• H2020 5GinFIRE OC2 – P4in5G (2019, 75.000 EUR) 
• H2020 MONROE OC2 – RECON (2017-2018, 100.000 EUR) 
• Ericsson Research project grants (2016-2023) 

5 important publications in the field

• Singh, S. K., Rothenberg, C. E., Langlet, J., Kassler, A., Vörös, P., Laki, S.and Pongrácz, G. (2022) ‘Hybrid P4 Programmable Pipelines for 5G gNodeB and User Plane Functions’, IEEE Transactions on Mobile Computing (IEEE TMC), pp. 1–18. (Q1, IF: 6.07)
• Fejes, F., Nádas, Sz., Gombos, G. and Laki, S. (2022) ‘DeepQoS: Core-Stateless Hierarchical QoS in Programmable Switches’, IEEE Transactions on Network and Service Management (IEEE TNSM), pp. 1–20. doi: 10.1109/TNSM.2022.3152017 (Q1, IF: 5.93)
• Laki, S., Györgyi, Cs., Pető, J., Vörös, P. and Szabó, G. (2022) ‘In-Network Velocity Control of Industrial Robot Arms’, 19th USENIX Symposium on Networked Systems Design and Implementation (NSDI'22), Renton, WA, USA, April 2022.
• Gombos, G., Kis, D., Tóthmérész, L., Király, T., Nádas, Sz. and Laki, S. (2022) ‘Flow Fairness with Core-Stateless Resource Sharing in Arbitrary Topology’, IEEE Access, 10, pp. 120312–12032. doi: 10.1109/ACCESS.2022.3222062 (Q1, OA, IF: 3.367)
• Györgyi, Cs., Laki, S. and Schmid, S. (2023) ‘P4RROT: Generating P4 Code for the Application Layer’, SIGCOMM Computer Communication Review (CCR), 53(1), pp 30–37. (Q2, IF: 2.322)