Why Does Satellite Latency Matter? – GEO vs LEO for Remote Sites

A satellite link can show a healthy speed test and still feel like wading through mud. Screens take a beat to refresh, SCADA polls crawl, VPNs time out and phone calls turn into polite interruptions. The culprit usually is not bandwidth. It is the time each signal spends traveling to space and back. So why does satellite latency matter so much, and how different are GEO and LEO links for a remote site?

Satellite latency matters because every request and reply has to travel to the satellite and back, and many remote-site tasks need dozens of those round trips in a row.

Geostationary (GEO) satellites orbit about 35,786 km (22,236 miles) up, so physics alone adds close to half a second to every round trip, and real links often land around 600 ms or more. Low Earth orbit (LEO) satellites fly roughly 500 to 1,200 km up, which cuts the delay to a few tens of milliseconds, close to a ground connection. GEO is fine for low-volume data that can wait. LEO behaves much more like cellular or wired broadband.

What Is Satellite Latency?

Latency is the time it takes data to get from one point to another. People usually measure it as round-trip time (RTT): how long it takes to send a small message and get an answer back. On a satellite link, a single request goes from your dish up to the satellite, down to a ground station, across the internet to a server, then back the same way.

Radio waves travel at the speed of light, about 299,792 km per second in a vacuum. That is fast, but at satellite distances it adds up. As the Satellite Signals latency explainer puts it, a single up-and-down hop through a geostationary satellite takes about 240 to 280 ms before any extra network delay.

How Far Away Are GEO, MEO and LEO Satellites?

Orbit height sets the minimum delay. The chart below shows the best-case round trip for each orbit, assuming the satellite is directly overhead and nothing else adds delay. Real links are slower, but the ratios hold.

Here is the math for the GEO bar. A round trip crosses the up-and-down path twice, so the signal travels four times the altitude:

• Distance: 4 × 35,786 km = 143,144 km
• Time: 143,144 ÷ 299,792 km per second = 0.4775 seconds, or about 478 ms

The same formula gives about 7 ms for a LEO satellite at 550 km. Real LEO links add routing, scheduling and ground station hops, so users typically see a few tens of milliseconds. GEO links add their own processing and congestion on top of the physics, which is how a 478 ms minimum becomes 600 ms or more.

Why Does Satellite Latency Matter More Than Bandwidth?

Because many jobs are chatty. They send a small request, wait for an answer, then send the next request. Bandwidth does nothing to shorten that wait.

SCADA Polling

A host that polls 50 devices one after another over a GEO link waits about 50 × 0.6 s = 30 seconds per scan in latency alone. Over LEO at 40 ms, the same scan spends about 2 seconds waiting. Alarms that depend on polling arrive later on GEO.

File Transfers and Web Traffic

Standard TCP, the protocol under most internet traffic, only sends a limited amount of data before it waits for an acknowledgment. With a classic 64 KB window and a 600 ms round trip, the most a single connection can move is about 64 KB ÷ 0.6 s, roughly 107 KB per second, or under 1 Mbps, no matter how big the satellite plan is. Modern window scaling and satellite acceleration help a lot, but the effect is why GEO links feel slow.

Voice and Remote Desktop

ITU-T guidance (G.114) recommends keeping one-way voice delay under about 150 ms for natural conversation. GEO cannot meet that, so calls feel awkward. Remote desktop sessions to an HMI lag behind every mouse click.

Security Handshakes

Setting up an encrypted VPN or a secure web session takes several round trips before any data moves. On GEO, that can add seconds to every new connection and cause timeouts in software tuned for fast networks. The answer is not to drop encryption. Good OT cybersecurity practice keeps tunnels encrypted, and engineers tune timeouts and keep tunnels up instead of rebuilding them constantly.

What Is Jitter, and Why Does It Matter Too?

Jitter is the variation in latency from one packet to the next. A link that always takes 600 ms is slow but predictable. A link that swings between 40 ms and 300 ms is harder for some software to handle, because voice calls stutter and protocols with tight timers may decide a device has gone missing.

LEO links can show more jitter than their low average suggests, since the terminal hands off between moving satellites and traffic may take different paths. GEO links get jittery when the shared capacity on a satellite beam is congested. Either way, watch the spread as well as the average when you test a link.

Does Weather Affect Satellite Latency?

Weather mostly affects signal strength rather than the travel time itself. Heavy rain absorbs and scatters radio signals, especially at the higher frequencies many satellite services use, an effect called rain fade. Snow and ice building up on a dish do similar damage.

When the signal weakens, the modem may switch to a slower, sturdier coding scheme, and lost packets have to be sent again. Each retransmission costs another round trip, so on GEO a storm can turn a sluggish link into an unusable one for a while. That is one more reason critical sites need a backup path that fails for different reasons.

How Do You Measure Latency on Your Own Link?

Start with a simple ping to a reliable server, repeated over a full day, so you see the busy evening hours and quiet nights. Record the average, the worst values and the packet loss. Many satellite modems and routers can log this continuously and alert you when latency climbs, which also helps prove a problem to the service provider.

GEO vs LEO for Remote Sites

GEO still has a place. It is mature, widely available and often fine for a site that sends small polls and alarms. But for anything interactive, LEO has changed the math for remote sites. Our guide to Starlink for oil and gas looks at how operators use LEO service in the field.

How Do You Work Around High Latency?

If GEO is your only option, or your backup, design for it instead of fighting it:

• Use report by exception, so sites speak up when something changes instead of waiting to be polled.
• Choose protocols that handle delay well, such as DNP3 with unsolicited reporting or MQTT.
• Poll multiple sites in parallel rather than one at a time where the system allows it.
• Use the TCP acceleration features built into most satellite modems.
• Process data at the site and send summaries rather than raw streams.
• Lengthen timeouts in SCADA and VPN settings so slow replies are not counted as failures.

Satellite Latency Is a Distance Problem You Design Around

Satellite latency comes from simple physics: signals moving at the speed of light still take time to travel tens of thousands of kilometers. GEO adds roughly half a second or more to every round trip, which slows chatty tasks like polling, VPN setup, voice and remote desktop. LEO cuts that to tens of milliseconds and behaves much more like a ground link.

So go through what your site actually does over its link and flag the chatty jobs: polling, VPN logins, voice and remote desktop. If several make the list and you are on GEO, either redesign them around the delay or price a LEO terminal along with the extra solar it will need.

← All guides