// /gcp-latency-check · 0 regions · live

GCP latency check.

Round-trip latency from your browser to Google Cloud regions. Hit in parallel, median ms reported. Click rerun for a fresh measurement.

GCP latency check — browser-measured RTT to Google Cloud regions using BigQuery regional endpoints.

Measured live from your browser to 0 GCP region endpoints. Three samples per region in parallel, fetch(no-cors) + median ms reported, four-second timeout.

GCP · 0 regions

measured from your browser · running… · sample size 3 · BigQuery endpoints
us-central1
Iowa
us-east1
S. Carolina
us-west1
Oregon
europe-west1
Belgium
asia-northeast1
Tokyo
asia-southeast1
Singapore

// endpoint used

<region>-bigquery.googleapis.com — the regional BigQuery frontend. Unauth GET / returns 404 quickly, but the TLS handshake still reaches the region-routed edge, which is what we time.

// what it measures

4 sequential samples per region; sample one is discarded as TLS warm-up and the median of the last 3 is reported. The figure approximates connection setup plus first byte — treat absolute values loosely and the ranking seriously.

// what it doesn't

Nothing is persisted and no backend ever sees your results. Region-to-region paths over Google's backbone aren't measured here — run mtr from a VM inside GCP for that.

// gcp latency faq

How does this GCP latency test work?

Your browser sends fetch(no-cors) probes to <region>-bigquery.googleapis.com, the regional BigQuery frontends. An unauthenticated GET just returns a 404, but the TCP and TLS handshake has to reach the region-routed frontend first, and that round trip is what gets timed. Four samples per region, the first discarded as warm-up, median of the remaining three displayed. No backend is involved and nothing is recorded — the whole measurement lives and dies inside your browser tab.

What is the difference between GCP's Premium and Standard network tiers?

This choice is unique to Google Cloud. Premium Tier (the default) ingests your traffic at the Google edge nearest the user and carries it over Google's private backbone to your region — fewer hops, lower jitter, higher egress price. Standard Tier does hot-potato routing: traffic crosses the public internet and only enters Google's network near the destination region, which is how most clouds behave anyway. If your latency budget is tight, Premium is usually worth it; for bulk egress, Standard saves real money.

How are GCP regions and zones named?

Google writes regions as geography plus compass plus number with no hyphen before the digit: us-central1, europe-west1, asia-northeast1 — note the difference from AWS's us-east-1 style, a constant source of typos in multi-cloud Terraform. Zones append a letter: us-central1-a through us-central1-f. us-central1 (Iowa) is GCP's historic heart, playing roughly the role us-east-1 plays for AWS: usually cheapest, broadest service coverage, and the default the console nudges you toward.

What does a global VPC mean for latency?

Unlike AWS and Azure, where a VPC or VNet is a regional object, a GCP VPC is a single global network with regional subnets. Two VMs on different continents in the same VPC talk over Google's backbone using internal IPs — no peering, no VPN, no public internet in the path. Cross-region internal latency still obeys physics, but it is consistent and private by default, and global internal load balancing with cross-region failover falls out of the design almost for free.

Does Cloud CDN change what users actually experience?

Yes. These numbers are browser-to-region, but GCP's global external Application Load Balancer hands you one anycast IP terminated at the Google edge nearest each user, and Cloud CDN caches responses right there. A user 180 ms from your region can get cache hits in under 30 ms. Treat this test as a guide for where to put origin compute and stateful services; for cacheable content sitting behind Cloud CDN, the region score matters far less.

Which GCP region should I choose?

Run this test from your users' geography and pick the consistent low scorer. Then sanity-check three GCP-specific angles: price (us-central1 and us-east1 are typically cheapest, while some Asian and South American regions cost notably more), multi-region storage (US, EU, and ASIA buckets and BigQuery datasets replicate across regions for you automatically), and sustainability — the console flags regions like europe-north1 with high carbon-free energy percentages if that matters to your organisation.

Why do my GCP numbers look flatter across regions than other clouds?

Google operates one of the largest private backbones on the planet and frequently picks traffic up at an edge near you before carrying it internally. Depending on how your ISP peers with Google, part of the journey to a distant region may run over Google's network rather than the public internet, compressing the differences between regions. The relative ordering in this test still holds, but the absolute gap between us-central1 and asia-northeast1 can look smaller than the equivalent comparison on another provider.

Need more network tools?
/netops has the rest.

open netops