Starlink is often described simply as satellite internet. In practice, delivering a working connection involves far more than a terminal communicating with a satellite overhead.
Data moves through a wider network made up of user terminals, satellites in low Earth orbit, optical links between spacecraft, ground stations and Points of Presence. Each component handles a different part of the journey between a user and the website, cloud platform or business system they are accessing.
Understanding that journey helps explain why Starlink performs differently from traditional satellite services, how it maintains connectivity while satellites move across the sky and where it fits within a modern enterprise network.
The journey from a Starlink terminal to the internet
A simplified Starlink connection looks like this:
The exact route can change from one moment to the next. The satellite serving the terminal will move, network demand will vary and traffic may return to Earth through different ground infrastructure.
The user sees a continuous internet connection. Behind the scenes, the network is constantly selecting and adjusting the route.
What does the Starlink terminal actually do?
Every Starlink connection begins with the terminal installed at the customer’s location.
Calling it a dish is convenient, but it can give the impression that it works like a traditional satellite antenna. Conventional satellite dishes are normally aligned towards a geostationary satellite that appears to remain fixed above the Earth. Starlink satellites move continuously relative to the user, so the terminal has a very different job.
It uses electronically steered phased-array technology to direct its signal towards available satellites. As one satellite moves out of range, the terminal prepares to connect to another. This transfer, known as a handover, allows an active session to continue while the underlying satellite connection changes.
A user could therefore remain on the same video call while the terminal communicates with several different satellites during the session.
Installation still matters
The terminal handles satellite tracking automatically, but its surroundings still affect performance. Trees, buildings, cranes and other obstructions can block sections of the sky used by the antenna.
This is particularly relevant for enterprise sites, where the easiest mounting location may not be the most effective one. A reliable deployment depends on selecting suitable hardware, checking for obstructions and considering the environment before installation begins.
Why does Starlink use low Earth orbit?
The main architectural difference between Starlink and traditional geostationary satellite services is the altitude of the satellites.
Geostationary satellites orbit approximately 36,000 kilometres above Earth, while much of the Starlink constellation operates at around 550 kilometres. The shorter journey between the terminal and the satellite helps reduce latency, creating a more responsive connection for applications such as video conferencing, cloud platforms and remote desktop access.
Low Earth orbit also changes how the network operates. A geostationary satellite appears fixed in the sky, allowing a traditional antenna to remain pointed towards one position. Starlink satellites move continuously relative to the user, so the terminal automatically hands the connection from one satellite to another as they pass overhead.
This requires a much larger constellation, but it also gives the network access to multiple satellites and routing options over time.
The trade-off is that a LEO network requires a much larger number of satellites. Each spacecraft covers a smaller area and remains visible to a particular terminal for a limited period.
Starlink maintains service by operating the satellites as a constellation rather than treating each one as a standalone connection.
What happens when the signal reaches the satellite?
Once the terminal sends data to a satellite, the network needs to determine how that traffic should continue towards its destination.
The satellite may send the data directly to a ground station within range. In other circumstances, it can pass the traffic to another satellite using an optical link.
The route depends on the position of the satellites, the destination of the traffic, the available ground infrastructure and current network conditions.
This means the familiar image of data travelling from a dish to one satellite and immediately back to Earth is only one possible route. Modern LEO architecture gives the network several ways to move traffic.
What are inter-satellite laser links?
Optical inter-satellite links allow Starlink satellites to communicate directly with one another using tightly focused beams of light.
One possible laser-linked route might therefore look like this:
Terminal → Satellite A → Satellite B → Satellite C → Ground station
This creates a connected network in space. Traffic can travel across the constellation before returning to Earth at a location with suitable ground infrastructure.
Laser links are especially useful over oceans, isolated regions and other areas where nearby ground stations may be limited. A satellite above one of these locations can pass traffic across the constellation until it reaches another satellite with access to an appropriate gateway.
Laser links give the network more routing options. They do not mean every connection always travels through several satellites.
Where a nearby ground station provides an efficient route, the network can send traffic back to Earth sooner. The benefit comes from having both options available.
Do laser links automatically mean lower latency?
Not necessarily.
Light travels faster through a vacuum than through fibre-optic cable, but overall latency depends on the complete distance and route taken. Sending traffic through several satellites could create a longer journey than using a nearby ground station connected to a strong terrestrial network.
The greater value of optical links is flexibility. They allow Starlink to extend service beyond the immediate reach of gateways, reduce dependency on a single terrestrial route and select different paths according to network conditions.
Latency is influenced by several other factors too, including the customer’s local Wi-Fi, routing after the traffic reaches Earth, network demand and the location of the destination server.
How do ground stations connect Starlink to terrestrial networks?
Data eventually needs to return to Earth. Ground stations provide the bridge between the satellite constellation and terrestrial communications infrastructure.
These facilities use multiple high-capacity antennas to communicate with satellites overhead. Once traffic reaches a gateway, it is carried over terrestrial backhaul to a Starlink point of presence
Ground stations perform a very different role from customer terminals. A terminal connects an individual location to the constellation. A gateway connects the constellation to high-capacity infrastructure on the ground.
Laser links reduce the need for a gateway to be close to every user, but they do not remove the need for gateways entirely. Most websites, cloud platforms, data centres and enterprise applications still operate on Earth, so satellite traffic eventually needs a route into terrestrial networks.
What is a Point of Presence?
After traffic returns to Earth, it still needs somewhere to enter the wider internet or an enterprise network. That connection takes place through a Point of Presence, commonly shortened to POP.
A POP is a network location, often based within a data centre, where Starlink connects with other infrastructure. It can provide onward access to internet providers, cloud services, content delivery networks and private enterprise environments.
Ground stations and POPs are sometimes confused, but they serve different purposes.
Ground station: moves traffic between satellites and terrestrial Starlink infrastructure.
Point of Presence: connects Starlink’s terrestrial infrastructure with the internet, cloud services or private networks.
For a standard internet connection, the POP is where traffic is handed towards its online destination. In more specialised enterprise architectures, it can also provide an interconnection point for private networking services.
Following one request through the network
Consider an engineer at a remote utility site opening a cloud-based monitoring platform.
The device first sends the request across the local network to the Starlink router and terminal. The terminal transmits it to a satellite passing overhead.
The Starlink network then routes the traffic over an available path. It might send the traffic directly to a ground station, or use one or more laser links to move it through the constellation first.t.
After reaching a ground station, the request enters Starlink’s terrestrial infrastructure and travels to a POP. From there, it is routed towards the cloud platform.
The response follows an appropriate route back through the network to the engineer.
All of this can happen in a fraction of a second, even though the connection may involve equipment in several locations and satellites moving thousands of kilometres per hour.
How does Starlink maintain a continuous connection?
A LEO satellite remains visible to a terminal for a limited period. Continuous service therefore depends on regular handovers between satellites.
The terminal and network coordinate this process as satellites enter and leave the usable part of the sky. A strong constellation provides overlapping coverage, allowing another spacecraft to take over before the existing link is lost.
Several factors work together to maintain service:
- Multiple satellites moving through view
- Automatic handovers by the terminal
- Alternative routes through laser links
- Access to different gateways
- Terrestrial routing through multiple POPs
This creates route diversity within the Starlink network, although it does not make every deployment immune to disruption. Local power, obstructions, damaged cables, installation quality and internal network design can all affect the service before traffic even reaches a satellite.
Why local network design still matters
A low-latency satellite connection can still deliver a poor user experience when the network around it has been badly designed.
Weak Wi-Fi coverage, congested access points, unsuitable routers and incorrect traffic management can all create performance issues that appear to come from Starlink.
For a single user at a small location, a basic setup may be enough. A business connecting dozens of devices, CCTV systems, cloud applications and operational technology will have very different requirements.
This is why enterprise Starlink deployments need to consider the complete network rather than focusing only on the terminal.
Where does Starlink fit within an enterprise network?
Starlink can operate as the main connection for a remote location, but that is only one possible use.
It can also provide temporary access while a fixed line is installed, backup for fibre or cellular connectivity, onboard service for a vessel or vehicle, or an additional path within an SD-WAN environment.
In each case, the Starlink network operates using the same underlying architecture. What changes is the service plan, hardware, installation and wider network design around it.
An office, passenger vessel and remote energy asset may all use Starlink, but they will rarely need the same configuration.
Putting the architecture together
The Starlink terminal gives the user access to the constellation.
Low Earth orbit satellites reduce the distance travelled by the radio signal. Handovers allow the connection to move between satellites, while laser links provide additional paths through space.
Ground stations bring traffic back into terrestrial infrastructure. POPs then provide the onward route into the internet, cloud platforms or private networks.
The performance experienced by the user is the result of all these components working together.
Final thoughts
The satellites receive most of the attention, but Starlink is better understood as a complete communications network.
Terminals, satellite handovers, laser links, ground stations, fibre routes and Points of Presence all contribute to moving data between a user and its destination. The route is dynamic, constantly changing as satellites move and the network responds to demand.
Understanding that architecture also helps businesses make better deployment decisions. The technology creates the connection, but hardware selection, installation, network design and ongoing management determine how effectively it supports the operation.
Clarus Networks designs, deploys and manages Starlink across remote sites, vehicles, vessels and multi-location enterprise networks. Our team works with organisations to determine where LEO connectivity fits and how it should integrate with their wider infrastructure.
