Banca de QUALIFICAÇÃO: CASSIUS CLAY BATISTA DA SILVA FILHO

Uma banca de QUALIFICAÇÃO de MESTRADO foi cadastrada pelo programa.
STUDENT : CASSIUS CLAY BATISTA DA SILVA FILHO
DATE: 18/05/2026
TIME: 08:30
LOCAL: meet.google.com/kgs-bsyw-trz
TITLE:

HyMRPL: A Hybrid and Mobility-Aware Extension of the RPL Protocol in IoT Environments

KEY WORDS:

Internet of Things; RPL; Wireless Sensor Networks; Hybrid Operating Modes; LLNs; 6LoWPAN.

PAGES: 97
BIG AREA: Ciências Exatas e da Terra
AREA: Ciência da Computação
SUBÁREA: Sistemas de Computação
SPECIALTY: Arquitetura de Sistemas de Computação
SUMMARY:

The Internet of Things (IoT) has driven the use of wireless sensor networks composed of devices with limited resources, unstable links, and high susceptibility to losses. In this context, the Routing Protocol for Low-Power and Lossy Networks (RPL) stands out as one of the main routing solutions for low-power and lossy networks (LLNs). However, RPL has structural limitations associated with the global definition of the Mode of Operation (MOP), which imposes the exclusive use of Storage or Non-Storing modes throughout the DODAG, restricting its flexibility in heterogeneous and dynamic scenarios. This work proposes, implements, and experimentally evaluates HyMRPL, a hybrid mode for RPL that allows the simultaneous coexistence of \textit{Storing} and \textit{Non-Storing} behaviors in the same DODAG, through local functional profiles (Class~S and Class~N) assigned individually to each node. HyMRPL uses the experimental value MOP~=~6, as permitted by RFC~6550, without altering the format of RPL messages and preserving full compatibility with the protocol specification. The proposal was implemented directly in the \texttt{rpld} daemon and validated in an emulation environment with Mininet-WiFi and 6LoWPAN. The experimental evaluation encompassed seven sets of experiments with topologies of 5 and 20 nodes, including static scenarios, mobility with real link degradation, fault isolation between classes, dynamic profile switching at runtime, automatic adaptive decision-making, scalability, and churn with simultaneous node entry and exit. The results demonstrate that HyMRPL maintains 100% PDR in static topologies, converges 35% faster than Storing mode with 20 nodes, and introduces no measurable overhead in memory, CPU, or control message volume. Class switching via POSIX FIFO operates non-disruptively, reversibly, and without emitting additional messages on the network. This work contributes to advancing the state of the art in routing for LLNs, offering a viable and flexible alternative for heterogeneous IoT applications.

COMMITTEE MEMBERS:
Presidente - 1874895 - RAMON DOS REIS FONTES
Interno - 2180207 - ITAMIR DE MORAIS BARROCA FILHO
Interno - 3139050 - ROGER KREUTZ IMMICH