Why Testing Methodology Matters in VPN Reviews
Most VPN review sites do not publish their testing methodology. This is a problem, because VPN performance varies enormously based on how and where you test it. A speed test run from a reviewer's home broadband in London will show completely different results than the same test run from a data center in Singapore on a 10 Gbps uplink. Marketing-funded reviews have a structural incentive to highlight favorable conditions; without knowing the test environment, there is no way to assess whether results reflect your use case.
GasVPN is funded through affiliate commissions. We disclose this clearly. What we control is the testing methodology: if we commit to running tests the same way for every provider, the relative rankings are valid regardless of the affiliate arrangement. This article explains exactly what we measure, how, and where.
Test Infrastructure
GasVPN maintains dedicated virtual machines co-located at five geographic points: New York (US East), London (EU West), Frankfurt (EU Central), Tokyo (Asia Pacific), and Sydney (Oceania). Each location runs on a host with a 1 Gbps symmetric uplink. We test from these locations rather than consumer broadband to eliminate ISP variability and ensure reproducible conditions across retests.
Speed tests use iPerf3 to a known-fast control server in each region — not public speed test websites, which can be affected by CDN optimizations that VPNs may or may not benefit from. We run three parallel iPerf3 streams to load the connection and report the median of 10 test runs per server location.
What We Measure
Throughput (Mbps): Measured on a 1 Gbps baseline with the VPN connected to the nearest same-country server. This gives the best-case speed number. We also test cross-continent routes (US East to Tokyo, London to Sydney) to assess long-haul performance.
Latency (ms): Round-trip time from the test machine to a fixed ping target (Cloudflare 1.1.1.1) with and without VPN connected. The delta represents the VPN's latency overhead.
Jitter: Measured over 100 consecutive pings during an active speed test to assess connection stability — relevant for gaming and video calls.
DNS leak test: We verify that all DNS queries resolve through the VPN provider's DNS servers, not the underlying ISP's, using dnsleaktest.com and our own DNS capture tooling.
IPv6 leak test: Separately confirmed via ipv6leak.com. Some VPNs pass the IPv4 DNS test but leak IPv6 DNS requests.
Kill switch test: We interrupt the VPN connection mid-transfer using iptables rules and verify that traffic stops immediately rather than routing over the unprotected connection.
Privacy and Jurisdiction Assessment
Speed is only one dimension. We evaluate each provider's legal jurisdiction (Five Eyes vs. Fourteen Eyes vs. neutral), no-logs policy, and whether it has been independently audited. Audit quality varies — a Cure53 or Deloitte audit with methodology published is different from an in-house policy document labeled "audited." We note audit providers, dates, and scope in each review.
Streaming and Torrenting
Streaming tests use separate residential IP addresses as a control and connect to Netflix, Disney+, BBC iPlayer, and Amazon Prime from each VPN provider's US, UK, and Australian servers. We test at least three separate servers per provider and note consistency across test days. Torrenting tests use a legal test torrent (Ubuntu ISO, well-seeded) and measure sustained download speed over 10 minutes.
Affiliate Transparency
GasVPN earns commissions when readers click affiliate links and purchase subscriptions. Rankings in our comparison tables are determined by test scores, not commission rates. Providers without affiliate programs appear in the database and receive the same testing treatment — they simply do not have "Get Deal" buttons. We do not accept payment for ranking position.
