Applied research division

Measuring the networks that
carry everything

Network Labs is the applied research group behind StreamMetrics. We study telemetry at the packet and flow level, model distributed-systems behavior, and build the instrumentation that makes large networks observable.

See our research areas Read publications
12+
Active research lines
38
Peer-reviewed papers
9
Open-source tools
2021
Founded
Research areas

What we work on

Our work spans the full path of a request — from the wire, through the kernel, across the mesh, to the dashboard.

Flow telemetry

Sampling strategies and lossless aggregation for high-rate network flows without overwhelming the collector.

Distributed tracing

Causal tracing across thousands of services with bounded overhead and accurate clock reconciliation.

Traffic modeling

Synthetic workload generation and failure simulation for capacity planning and resilience testing.

Observability UX

How engineers actually read dashboards — visual encodings that reduce mean-time-to-insight under pressure.

Telemetry integrity

Tamper-evidence and provenance for metric pipelines, so dashboards can be trusted during incidents.

Edge instrumentation

Lightweight agents for constrained edge nodes that report without degrading the workload they observe.

Publications

Selected recent work

A sample of papers and technical reports from the lab.

2026-02
Bounded-Overhead Causal Tracing in Service Meshes
Proc. of NSDI ’26 · 14 pp.
2025-11
Lossless Flow Aggregation at Line Rate with Adaptive Sketches
ACM SIGCOMM CCR · 11 pp.
2025-08
A Field Study of How SREs Read Latency Dashboards
Technical Report NL-TR-2025-04 · 22 pp.
2025-05
Provenance Tags for Trustworthy Metric Pipelines
IEEE/IFIP DSN ’25 · 12 pp.
2025-01
Synthetic Traffic That Breaks the Way Real Traffic Breaks
Technical Report NL-TR-2025-01 · 18 pp.
About

A small lab with a long horizon

Network Labs operates as the independent research division of StreamMetrics. We publish openly, release tools as open source, and partner with university groups on questions that won’t pay off for years — because the networks we all depend on deserve that kind of attention.

Research collaboration & press: research@network-labs.example