Design and performance analysis of a border router architecture and its digital twin with support for low latency low loss scalable traffic

• Grant number: 23/00420-1
• Support Opportunities: Regular Research Grants
• Start date: October 01, 2023
• End date: September 30, 2025
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Agreement: MCTI/MC
• Principal Investigator: Djamel Fawzi Hadj Sadok
• Grantee: Djamel Fawzi Hadj Sadok
• Host Institution: Centro de Informática (CIn). Universidade Federal de Pernambuco (UFPE). Ministério da Educação (Brasil). Recife , SP, Brazil
• Associated researchers: Judith Kelner
• Associated scholarship(s): 24/16293-1 - Development of the management and telemetry module for L4S transport and forwarding protocols, BP.TT; 24/15650-5 - Experimental evaluation of data synchronization between NDT and the real world in a L4S-IoT environment and adaptive optimization, BP.TT; 24/15604-3 - Development of the L4S-IoT NDT telemetry and synchronization module., BP.TT; + associated scholarships 24/03159-5 - Design of a router with support for L4S and its Network Digital Twin, BP.TT; 23/15430-2 - Design of a router with support for L4S and its Network Digital Twin, BP.TT; 23/13631-0 - Adapt and test transport protocols with L4S support to ARM devices and data modelling, BP.TT; 23/12895-4 - Design of discovery modules, Linux data plane monitoring and Network Digital Twin optimization, BP.TT; 23/13433-4 - Linux and NDT data plan optimization, BP.TT - associated scholarships

Abstract

L4S-IoT develops a new router architecture with support for IoT traffic, considering the lack of solutions at the network edge that ensure its QoS requirements. The project seeks a solution that provisions TCP/IP networks with hardware supporting mechanisms that protect specific IoT flows: those with low latency and low packet loss requirements. The proposed architecture combines the design of new hardware, software, with protocols and is based on the following three main recent technological advances:a) Advances in the optimization of the Linux data plane using packet processing techniques with better performance;b) Use of Low Latency Low Loss Scalable throughput (L4S) architecture to guarantee low losses and latencies to specific IoT flows and;c) Modeling and optimization of the L4S-IoT use case based on Network Digital Twin (NDT).The proposal is innovative considering that:a) There are no edge routers on the market that support the L4S architecture;b) To the best of our knowledge, it is the first time that the L4S framework is integrated with data plane optimization technologies including DPDK and SR-IOV to handle IoT traffic;c) The solution will use "bare metal" hardware, avoiding dependence on proprietary network operating systems (NOS) and;d) The lack of a Digital Twin representing a scalable IoT service network with low latency and packet loss guarantees.The adopted solution must scale to support:a) Large IoT flows as in the case of video streams generated by traffic monitoring cameras or/or public safety in smart cities and;b) A massive number of low-power IoT device flows.The resulting product of LS4-IoT includes:a) The development of an edge router with support for L4S architecture;b) Design and evaluation of Linux data plane optimization algorithms (for packet processing) using technologies such as DPDK and/or SR-IOV;c) Research, specification and development of a Digital Twin Network (DTN) of the L4S-IoT scenario and the;d) Development of an L4S-IoT service management application capable of:a. Automatically detecting when problems occur in ECN support on routers;b. Checking when routers are resetting the ECN bits;c. Verifying when some servers do not reduce their transmission rates as expected.Although the proposed solution is that of an edge router, it may be extended, in the future, to other network levels such as aggregation and core or backbone networks.The L4S-IoT solution is based on the following steps:a) Installation and configuration of a network operating system on a "bare-metal" router (a generic hardware device without software);b) Research, definition and specification of the L4S architecture supporting data plane optimization;c) Development of the L4S software for the chosen device (network appliance);d) Specification and evaluation of an LS4-IoT service management application;e) Conducting tests and evaluating the performance of the router in a controlled testbed environment;f) Specification, development and use case analysis using the Digital Twin for L4S-IoT.In summary, the project develops an L4S architecture for IoT services. A first core task is the development of a router with L4S support. This is accomplished using a bare metal appliance, installing an open NOS, and developing appropriate active queue management (AQM) policies. Next, different Linux data plane optimization strategies are evaluated. Algorithms based on learning techniques will be developed for the discovery, monitoring and optimization of this plan. Additionally, the project will design an NDT capable of detecting anomalies, predicting events, allowing future planning, in addition to suggesting new configurations. (AU)

Scientific publications

(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)

FILHO, ASSIS T. DE OLIVEIRA; FREITAS, EDUARDO; DO CARMO, PEDRO R. X.; SOUTO, EDUARDO; KELNER, JUDITH; SADOK, DJAMEL F. H.. Analysis of SR-IOV in Docker containers using RTT measurements. COMPUTER COMMUNICATIONS, v. 228, p. 15-pg., 2024-10-05. (23/00420-1)