Globalping and RIPE Atlas are the two networks people reach for when they need to measure the internet from many places at once. From a distance they look alike. They were built for different work.
- Pick RIPE Atlas when you need the widest reach and open data. It has by far the most probes of the three, and it has been running since 2010.
- Pick Globalping when you need an answer now. A test needs no account, and it runs from a terminal, a chat app or a browser.
- Look at Hopbits when the question is whether a service works for people in a country, checked on a schedule, with an alert when it stops.
Hopbits is ours, and it is the youngest and smallest of the three. This page says where each of the other two is the better choice.
Side by side
| Globalping | RIPE Atlas | Hopbits | |
|---|---|---|---|
| The network | |||
| Run by | The jsDelivr teamLive since May 2023. | The RIPE NCCA not-for-profit internet registry. Since 2010. | HopbitsA company in Lithuania. Since 2026. |
| Size | 5,597 probesIn 124 countries and 1,090 cities, by its own count. | 13,421 probesConnected in February 2026, with 974 anchors. Devices in 183 countries. | Far smallerIt is young. The status page counts its nodes live. |
| Hosted by | Anyone, in a containerSponsors can get a hardware probe. Globalping runs its own in key places. | VolunteersAs a small device, as software, or as an anchor. | Independent operatorsOn home, mobile and datacenter lines. They earn points. |
| What it does | |||
| Measures | Ping, traceroute, DNS, MTR, HTTP | Ping, traceroute, DNS, TLS, NTP, HTTPHTTP only towards its own anchors. | HTTP, TCP, ping, DNS, traceroute, MTRAnd where an IP address is. |
| Monitoring with alerts | No: Not built inTests on demand. Other tools add the schedule. | Partly: Status checksRepeating measurements can be a basis for alerts. | Yes: Built inOn a schedule, with alerts by email and webhook. |
| Using it | |||
| A test without an account | Yes: Yes250 tests an hour, 500 when signed in. | No: Account neededIt is free. Reading the data needs none. | No: Account neededIt is free. |
| What it costs | FreeMore tests by hosting a probe or sponsoring. | CreditsA starting grant, then earned by hosting a probe. RIPE NCC members get them monthly. | Credits1,000 free every day. Packages with more are paid. |
| Who sees the results | EveryoneResults are public, and kept for up to about six months. | Open dataMost of what it collects is free for anyone to read. | YouIn your account and through your API keys. |
| Open source | Yes: YesBy its own description. Probe, platform and CLI are on GitHub. | Partly: The software probeUnder GPL-3.0. | Partly: Node client and CLIOn GitHub under Apache-2.0. The service that hands out jobs is not. |
When RIPE Atlas is the right choice
RIPE Atlas is the oldest and the largest of the three, and it is run by a not-for-profit internet registry.
- Reach. More than 13,000 connected probes, with devices in 183 countries.
- Open data. Most of what it collects is free for anyone to read. That is what research needs.
- Anchors. Beside the probes it has close to a thousand anchors, which are also the targets its HTTP measurements go to.
What it asks of you: an account, and credits. A new account can claim a starting grant. After that, the usual way to earn credits is to host a probe.
When Globalping is the right choice
Globalping is the quick one. It is made for the moment a route looks wrong and you want to see it from twenty places at once.
- No account for a test. Open the site or the command-line tool and run it.
- Where you already work. It has a command-line tool, apps for Slack and Discord, and an API.
- In the open. It describes itself as an open-source project, and its code is on GitHub.
What it leaves to you: watching a service over time. It runs the test you ask for. The schedule and the alert are yours to build.
Where Hopbits fits
Hopbits starts from a different question. The other two ask what the network looks like from a place. Hopbits asks whether your service works for the people there, and tells you when it stops.
- Monitoring is the product. A monitor checks on a schedule from the countries you choose and alerts you by email or webhook. You do not build that part.
- Several nodes answer, and the answers are compared. One bad line does not pass for an outage.
- Operators earn points for staying online. The aim is nodes on home and mobile connections, in the places a data centre does not reach. The goal is in the litepaper.
- It checks where an IP address is. IP location times replies from several nodes and says how sure the answer is.
What it cannot give you yet is size. The network is new, many countries have no node, and the other two have years of data that Hopbits does not. If you need a view from a particular country today, look at the status page first.
They work together
These are not rivals for one job. A team can watch its service with a monitor, reach for Globalping when a route looks wrong, and read RIPE Atlas when it wants to know how a whole network behaved last month.
Sources
What this page says about other services was read on their own pages on 22 November 2026. Things change. If a line here is out of date, write to [email protected] and we will correct it.
- About Globalping: the network in numbersGlobalping
- Credits and limitsGlobalping
- Globalping: measurement types, results and integrationsjsDelivr on GitHub
- The Globalping probejsDelivr on GitHub
- Globalping goes liveGlobalping blog, 18 May 2023
- A day in the life of RIPE AtlasRIPE Labs, 3 March 2026
- RIPE AtlasRIPE NCC
- What is RIPE Atlas?RIPE Atlas docs
- The credit systemRIPE Atlas docs
- User-defined measurementsRIPE Atlas docs
- Status checksRIPE Atlas docs
- What we doRIPE NCC
- The RIPE Atlas software probeRIPE NCC on GitHub
