Projects
Research & professional
-
SKA-Low Correlator & Beamformer
I work across several aspects of the Correlator and Beamformer (CBF) — the system that combines signals from thousands of antennas — for SKA-Low, the low-frequency telescope of the Square Kilometre Array, one of the world's largest radio-astronomy projects.
- Software-Defined Networking — lead the SDN design and architecture for the CBF.
- Validation — lead CBF validation, including the Factory Acceptance Tests (FAT).
- System integration — co-own the architectural decisions for integrating the CBF with the rest of the telescope, especially the downstream interface to Science Data Processing (SDP).
- Control & management — contribute to the Tango-based control and management framework for the CBF.
- Radio astronomy
- SKA-Low
- Software-Defined Networking
- Correlator & Beamformer
- System integration
- Tango Controls
-
CryoPAF
Lead designer & architect — SDN
I was the lead designer and architect of the Software-Defined Networking for CryoPAF, CSIRO/ATNF's cryogenically-cooled phased array feed (PAF) receiver for radio astronomy.
- Radio astronomy
- Phased Array Feed
- Software-Defined Networking
-
Quasar Satellite Technologies
Lead designer & architect — SDN
I was the lead designer and architect of the Software-Defined Networking for Quasar Satellite Technologies, developing phased-array ground-station technology for satellite communications.
- Satellite communications
- Phased array
- Software-Defined Networking
-
SDN systems research
Chief Investigator & student supervision
Research on the systems side of Software-Defined Networking: where to place controllers at scale, how to garbage-collect stale forwarding rules, and how to build high-performance, protected network stacks — including DLibOS, a per-core stack presented at ACM ASPLOS.
- Software-Defined Networking
- Distributed systems
- Controller placement
- High-performance networking
-
SDN for Time-Sensitive Networking
Co-investigator
A funded project bringing Software-Defined Networking to Time-Sensitive Networking (TSN) — programmable control of deterministic, low-latency Ethernet for industrial and real-time systems.
- Software-Defined Networking
- Time-Sensitive Networking
- Deterministic networking
-
Network verification & provenance
Chief / co-investigator
Research on making networks verifiable and accountable: formally verifying control protocols with timed automata, and tracking the provenance of network knowledge so configuration and state can be explained and trusted.
- Formal verification
- Timed automata
- Network provenance
- Network management
-
Deep Bypass
Product Owner
Tools for profiling modern enterprise-class networks — combining deep learning with kernel-bypass packet capture to classify encrypted traffic and discover network assets in real time at 40 Gbps and beyond.
- Deep learning
- Encrypted traffic
- High-speed packet capture
- Traffic classification
-
Synthetic network traffic generation
Co-investigator & student supervision
Generating realistic synthetic network traffic with deep generative models (GANs, diffusion, and time-series imaging) to support measurement, testing and privacy where real traces are scarce or sensitive.
- Generative models
- Synthetic data
- Network measurement
- Deep learning
-
Combatting cyber-fraud
Co-investigator & student supervision
Detecting and characterising fraud in online social networks — using honeypot accounts and behavioural analysis to expose fake-engagement ("like farm") ecosystems and protect advertisers.
- Fraud detection
- Social networks
- Honeypots
- Security
-
Private real-time processing of outsourced functions
Co-investigator & student supervision
Privacy-preserving outsourcing of network functions to the cloud — cryptographic techniques (e.g. SplitBox) that let a provider run functions such as firewalls without learning their rules, policies or data, while keeping real-time performance.
- Cloud security
- Network Function Virtualization
- Applied cryptography
- Privacy
-
Internet-wide measurement platform
Co-investigator & student supervision
Large-scale active Internet measurement — scanning to study DNS, TLS certificates and protocol deployment worldwide, informing routing, content delivery and network security.
- Internet measurement
- DNS
- TLS
- Network security
-
Distributed delay-tolerant payments with blockchains
Co-investigator & student supervision
A blockchain-based payment scheme for regions with poor connectivity: miners handle everyday transactions peer-to-peer while a bank joins only periodically for currency exchange and oversight — cutting cost without losing security.
- Blockchain
- Delay-tolerant networking
- Payments
- Distributed systems
-
Blockchain consensus & security
Co-investigator & student supervision
Research on the security and performance of blockchain systems: quantifying double-spending risk and demonstrating consensus attacks (e.g. against proof-of-authority) under realistic network delays, across public and consortium blockchains.
- Blockchain
- Distributed systems
- Security
- Consensus
-
FORGE
Contributor
An EU project bringing remote experimentation on FIRE testbeds into eLearning — building tools and interactive learning resources that let students run real network experiments from within course material.
- eLearning
- FIRE testbeds
- Remote experimentation
-
OMF / OML / LabWiki
Core contributor
A toolchain for reproducible networking experiments on testbeds: OMF for experiment control and management, OML for measurement collection, and LabWiki for orchestrating and documenting repeatable experiments.
- Networking testbeds
- Reproducible research
- Measurement
Personal projects
-
AstroForecast
A cross-platform app that, for any location, gives a 3-night astronomy "seeing" forecast — a home-grown observability estimate built from free Open-Meteo weather data — plus an aurora forecast (current probability from NOAA SWPC's OVATION model, with Kp context). Aurora alerts near a saved location are planned.
- Cross-platform app
- Astronomy
- Open-Meteo
- NOAA SWPC
-
CICD-RAG
A training-free, fully local RAG system that classifies CI/CD failures and suggests mitigations. Retrieval is hybrid — dense (nomic-embed-text) plus sparse (BM25) over a Qdrant collection — and classification is a structured-output call to a local Ollama model.
- RAG
- CI/CD
- Qdrant
- Ollama
- BM25
- Local LLM