Workstations, test devices and data-collection tools for research. Contact us to discuss availability and access.
Multicore research workstations support data preparation, feature extraction, simulation and classical machine learning, with GPU acceleration for lightweight inference and small-model experiments.
Access to shared departmental workstations can be coordinated for approved projects. The facilities support database systems, machine learning, deep learning and network-security experimentation.
Collection and preparation tools distinguish original traffic, imported benchmarks and synthetic examples. Java and Python components support feature extraction, classical machine-learning and neural-network experiments, split checks and experiment records. The public Android companion uses the PCAPdroid capture engine.
Explore the public collector ↗Research inputs are tracked as original collection, imported benchmarks or synthetic examples. Derived features, split assignments and experiment records support analysis and traceability. See the encrypted-traffic categories below for availability and access.
Controlled Android test devices, including rooted-device setups where a study requires them, support approved instrumentation and traffic-collection experiments. Device requirements and access are agreed for each project.
Compute, data and device access are allocated to approved projects. Additional capacity depends on the project’s requirements and funding.
Four categories describe where the research inputs come from and how they are prepared.
Traffic contributed by Ayush Kumar and volunteers during normal device use, without prescribed application tasks. The study has documented consent and institutional ethics approval.
A truncated public edition is prepared locally and has not been published. Requests for raw research access from trusted researchers and organisations are reviewed for the proposed study.
Enquire about research access →The research also uses ISCXVPN2016 material from the University of New Brunswick’s Canadian Institute for Cybersecurity. This is an external benchmark, kept separate from the team’s original collection.
Use the original provider’s dataset page for access information and the required citation.
View UNB/CIC’s dataset page ↗Clearly labelled generated examples support experimental input preparation and pipeline checks. They are kept separate from observations of volunteers.
Requirements and availability are reviewed for the proposed study.
Discuss your study →Prepared features, split assignments, input and script hashes, code revisions and runtime records support analysis and traceability. These artifacts remain derivatives of their source data.
Tell us which features, data sources or evaluation support your study needs.
Discuss data and evaluation support →Tell us your research question and the data, software or evaluation support you need. The team reviews access against the study’s consent, ethics, licensing and privacy requirements.
Tell us about your project, software, data size, compute needs and timeline. We’ll check what is available.
Shared facilities are scheduled, not exclusively allocated. Dataset sharing and additional capacity are confirmed for each approved project.