Provider × provider in Europe
From your country: the fastest link from a data centre there to the other provider's network – switch the country below. Alternatively the best value or the median of all location pairs in Europe. On the right: how close each EU cloud is to the nearest other EU cloud and to the nearest hyperscaler.
As soon as the first measurement servers report values, the analysis appears here.
The European mesh
The rings show the planned measurement locations in Europe. As soon as the probes run, the measured connections appear here – coloured from fast to slow.
Providers and locations
What you need it for
EU cloud plus a hyperscaler service
Keeping data in a European cloud and using single services of a hyperscaler means every request pays the latency between the two. How close the data centres are decides whether that works.
Replication and fallback site
Database replication, backup or a second provider for emergencies: synchronous only works at a few milliseconds. Cloud Mesh shows which European providers are close enough to each other for that.
Staying in Europe without guessing
Sovereignty is a matter of provider, law and operations – latency is physics. Both belong in the decision; Cloud Mesh supplies the measured side.
How Cloud Mesh works
- 1
A small measurement server runs in the data centres of the providers – from Frankfurt via Paris and Amsterdam to Stockholm and Milan.
- 2
Every three minutes every location measures every other: ping and TCP connection set-up, the median per direction.
- 3
Once a month this becomes the official matrix; in between, the values of permanently running probes flow in.
How we measure
- Per direction the median (p50) counts; per location pair the mean of both directions. A direction is valid with at least 10 replies and at most 50 % loss – if a provider blocks ping, the TCP handshake is used instead.
- The physical minimum is the travel time in fibre: light covers about 204 km per millisecond in glass. The factor shows how far a connection lies above it – detours in the network, peering, routing.
- We measure data centre to data centre. The latency to your users also depends on their access network.
- We measure in Europe: the hyperscalers’ European regions and the data centres of the European providers, including Switzerland, the UK and Norway.
Data status & download
No measurement published yet.
- Current matrix as JSON (cloud-mesh.matrix/v1)
- All published measurements
- Full methodology of comparison.cloud
Planning an architecture across several clouds?
In the Cloud Assessment of the TIMETOACT GROUP, measured latency and list prices feed into the target picture and the provider choice. Or ask Polaris directly.
Frequently asked questions about Cloud Mesh
What is Cloud Mesh?
Cloud Mesh is the latency comparison of comparison.cloud: measurement servers in the European data centres of the cloud providers continuously measure each other. The page shows how close the EU clouds are to each other and to the hyperscalers’ European regions – as a map, provider matrix and location table.
Why does Cloud Mesh only measure in Europe?
comparison.cloud compares clouds for organisations in Europe. That is why the probes run in European data centres: at the European providers and in the European regions of the hyperscalers – including Switzerland, the UK and Norway. Regions outside Europe are not measured.
Which European cloud is closest to AWS, Azure or Google?
The table “How close the EU clouds are to each other” shows it: for every European provider the measured latency to the nearest hyperscaler data centre in Europe and which hyperscaler that mostly is. The values are measured, not estimated from distances.
Are a few milliseconds enough for synchronous replication between two providers?
That depends on the application. As a rough guide, synchronous database replication usually needs single-digit millisecond latency, while asynchronous replication and backup tolerate considerably more. Cloud Mesh supplies the measured latency – assessing it for your architecture belongs in the assessment.
Is low latency a sign of a sovereign cloud?
No. Latency says something about the physical proximity of data centres and the quality of networks, nothing about legal framework, ownership or operations. For sovereignty there is the Sovereign Compass; Cloud Mesh adds the measured network side.