Research
Research directions
My work focuses on networked systems engineering across protocols, transport technologies, and evaluation methods, with an emphasis on timing, latency, predictability, and system behavior under real operating constraints. I organize it around a small set of broad directions so that projects and publications map back to recurring technical themes.
Clock synchronization
Uncertainty analysis, timestamping behavior, and distributed timing quality.
02Deterministic networking & latency modeling
Predictable communication, bounded latency, and timing-aware system design.
03Digital twins
System representations for analysis, validation, and experimentation.
04Tree splitting algorithms
Algorithmic approaches to contention resolution and scalable coordination.
01 / Timing systems
Clock synchronization
Clock synchronization remains a core theme in my work. I study how synchronization uncertainty can be quantified and across systems and network paths, with attention to timestamping, queueing, switch behavior, oscillator stability.
I spend a lot of effort in mathematically modeling clock synchronization protocols like PTP and NTP and verifying the models with measurements. The work has resulted in four publications, with more in progress and a patent pending.
02 / Networking systems
Deterministic networking & latency modeling
Related closely to my work on clock synchronization, I also model and measure latency and jitter in networked systems.
Many factors affect the latency and jitter of networked systems, including scheduling policies, queueing behavior, and the interaction between network traffic. Similar to my work on clock synchronization, I build models and verify them with measurements.
03 / Models & evaluation
Digital twins
Accurate modeling of systems has also led me to explore Digital Twins - keeping live representations of systems that can be used for analysis, validation, and experimentation.
This direction connects naturally to system evaluation, what-if analysis, and the translation between observed behavior and the abstractions used to reason about it. The work also involved looking at cloud service behaviour, especially on Microsoft Azure.
04 / Algorithms
Tree splitting algorithms
During my reasearch internship during my masters, I worked on the mathematically beautiful problem of tree splitting algorithms in contention resolution.
I built models and simulations to investigate and validate the latency, fairness and efficiency of tree splitting algorithms in various contention scenarios. The work has resulted in various publications.
Methods
How I approach research
Model
Develop analytical, stochastic, and system-level models that expose where timing and performance behavior come from.
Measure
Validate assumptions with traces, experiments, prototypes, and implementation-aware measurements.
Connect
Map projects and publications back to broader directions so individual results contribute to a coherent body of work.