Netperf vs Netpicker

Netperf is an open-source, self-hosted utility for measuring network performance. It reports unidirectional throughput and end-to-end latency through a client-server arrangement: netperf runs tests and netserver supplies the server side. Netserver can run as an inetd child or standalone daemon. The manual describes TCP and UDP transfer and request/response tests over IPv4 and IPv6 sockets, plus XTI, DLPI, Unix domain socket, and SCTP tests. Where the platform and build support it, results can include CPU utilization and service demand. Optional configuration enables STREAM pacing, timing histograms, and interim results. The primary distribution is source code, intended for systems without readily available binaries. Reported Windows builds use Cygwin, the Windows DDK, or Microsoft Visual Studio 2003; the README discusses an OSX naming workaround. Hewlett Packard Enterprise released this version under the MIT license. The project warns that its manual may be outdated, and several test families require explicit configuration.

Netpicker validates network state, device configurations, compliance, security and behavior, and can automate remediation. It runs automated Python tests in pytest format against device configurations, show-command output and API data. The free version includes unlimited configuration backups, Git history, diffs, search and drift detection. Netpicker supports devices from more than 180 vendors and can be deployed as a Docker container or OVA virtual appliance, on-premises or in air-gapped environments. API-based connections link it to source-of-truth systems, Git, ITSM tools and other existing tools. Its REST API manages devices, backups and tests, runs show commands and tests, and restores backups; documentation is provided through Swagger/OpenAPI. Security validation covers CVEs, security configurations, compliance and pen-test findings without sending network data outside the customer's environment. Potential CVEs are matched by OS version, while verified CVEs also check whether vulnerable features are enabled and mitigations are present. AI can draft tests, remediation jobs and troubleshooting queries from plain English, subject to engineer review and approval. The free Foundation plan is available; Professional starts at $7,500 a year.