Chapter 2
Satellite internet has to solve physical problems that terrestrial networks never face. The user's "tower" moves at orbital speed. The signal travels thousands of kilometers before it reaches a router. A constellation may contain tens of thousands of nodes that must be kept coordinated, licensed and, at end of life, safely retired. WIA-COMM-005 treats these as one connected problem set. Its stated aim is global broadband for remote, rural, maritime and aviation users, which the standard frames with the principle 弘益人間 (Benefit All Humanity). This chapter walks through the main difficulties in the order a signal meets them: distance, link budget, motion, the space and ground segments, and finally regulation.
Round-trip latency is set largely by orbital altitude, and no protocol can remove it. The standard's Phase 1 data-format document describes three orbit classes, and the gap between them is the starting point for most design decisions.
| Class | Altitude | Round-trip latency | Typical use |
|---|---|---|---|
| LEO | 500–2000 km | 20–40 ms | Low-latency user access |
| MEO | 8,000–20,000 km in the architecture section (the terminology list gives 2000–35,786 km) | 100–150 ms | Navigation, regional broadband, LEO backup |
| GEO | 35,786 km | 500–600 ms | TV broadcast, legacy maritime and aviation, weather, emergency backup |
GEO needs only three to five satellites for global coverage and the antenna can stay pointed at one spot. The price is a half-second round trip that interactive applications feel immediately. LEO reaches 20–40 ms, but a single satellite sees a small patch of ground and passes quickly overhead, so continuous service requires a large constellation. The typical configuration in the specification uses altitudes of 550 km (Starlink), 600 km (OneWeb) and 590 km (Kuiper), 20–80 orbital planes, 20–80 satellites per plane, and totals from 300 to more than 40,000 satellites. The orbital period is roughly 90–100 minutes.
The standard does not pick a winner. It describes hybrid architectures in which LEO carries latency-sensitive user traffic, MEO fills regional coverage, and GEO handles broadcast and backup. The challenge is then coordination across orbits: one terminal and one operator view have to span systems with very different delay and visibility behavior.
Inclination adds a second trade-off. Polar inclinations of 87–90° cover the entire globe including the poles, which suits scientific, military and emergency use. Inclinations of 53–70° concentrate capacity on population centers, which suits commercial broadband. A constellation cannot be optimized for both at once, so the choice of inclination is also a choice of who gets served well.
Every link, user or gateway, depends on whether received signal power clears the noise floor by enough margin. The standard gives the link budget as four relations:
EIRP (dBW) = Tx Power (dBW) + Antenna Gain (dBi) - Losses (dB)
Path Loss (dB) = 20 log10(d) + 20 log10(f) + 92.45
SNR (dB) = EIRP - Path Loss - Noise Figure + G/T
Capacity (Mbps) = Bandwidth × log2(1 + SNR)Here d is distance in km, f is frequency in GHz, and G/T is the receiver's gain-to-noise-temperature ratio. The path-loss term shows why orbit choice matters so much. Taking the 28 GHz Ka-band example used elsewhere in the standard, a 600 km path gives about 20·log10(600) + 20·log10(28) + 92.45 ≈ 177 dB. A GEO path of 35,786 km at the same frequency gives about 212 dB. That is roughly 35 dB more loss, or a factor of more than three thousand in power, before any weather or pointing error. These figures are nadir-distance illustrations. Real slant ranges are longer, and the gap persists.
Two practical cautions come with the capacity formula. First, the Shannon expression needs SNR as a linear ratio, not in decibels, so an implementation must convert before applying log2(1 + SNR). Second, the result is an upper bound for the given bandwidth and signal quality. It is not a throughput guarantee, because real modems, coding and protocol overhead fall short of it.
Frequency choice is the other half of the budget. The terminology section defines the bands the standard works with:
Higher bands offer more spectrum, and therefore more bandwidth in the capacity formula, and they allow smaller antenna apertures for a given beamwidth. Path loss grows with frequency through the 20 log10(f) term. The standard's SDK example makes this concrete with a Ka-band link at 28 GHz and 250 MHz of bandwidth, a 10 W transmitter and a 35 dBi antenna over 600 km.
const linkBudget = calculateLinkBudget({
distance: 600, // km
frequency: 28, // GHz (Ka-band)
txPower: 10, // W
antennaGain: 35, // dBi
bandwidth: 250 // MHz
});
console.log(`SNR: ${linkBudget.snr.toFixed(2)} dB`);
console.log(`Capacity: ${linkBudget.capacity.toFixed(2)} Mbps`);The same calculation is available from the command line as wia-comm-005 link-budget --distance 600 --frequency 28 --power 10. Having both entry points lets an engineer check a hand calculation against the library's output.
A LEO satellite moves at about 7.5 km/s relative to the ground. That speed causes two problems the standard treats explicitly: frequency offset and short visibility windows.
Doppler. The compensation relation is Δf = (v/c) × f₀, with c given as 299,792 km/s. For v = 7.5 km/s the ratio v/c is about 2.5 × 10⁻⁵. At a 28 GHz carrier that is on the order of 700 kHz. This is an upper bound for motion directly along the line of sight, and the actual shift varies with geometry during a pass. A receiver that expects a fixed carrier will not lock, so terminals and gateways must predict and remove the shift continuously. The command wia-comm-005 doppler --velocity 7500 --frequency 28000 reproduces the calculation.
Visibility. A satellite is usable only above a minimum elevation angle, typically 25–40° according to the standard. The specification expresses the geometry as follows:
cos(θ) = R_e / (R_e + h)
r = R_e × arccos[R_e × cos(θ) / (R_e + h)] // coverage radius at minimum elevationwith R_e = 6,371 km. Raising the minimum elevation angle shrinks each satellite's footprint, which improves signal quality because paths are shorter and cross less atmosphere. Smaller footprints also mean more satellites are needed for continuous coverage. This is a direct cost trade-off built into constellation sizing.
Handover. Because a satellite crosses the sky within minutes, a user session outlives many satellites. The standard gives the prediction relation as t_handover = arccos(R_e / (R_e + h)) / ω, where ω is the satellite's angular velocity. For a 550 km altitude and a period near 95 minutes, this evaluates to a few minutes. That is consistent with the idea that a terminal must hand off many times during a single long download. The SDK exposes this as predictHandover, which takes a current satellite identifier, the user's latitude and longitude, and an elevation angle, and returns the time remaining and the next satellite.
const handover = predictHandover({
currentSatelliteId: 'SAT-1234',
userLatitude: 37.7749,
userLongitude: -122.4194,
elevationAngle: 25
});
console.log(`Next handover in: ${handover.timeRemaining} seconds`);
console.log(`Next satellite: ${handover.nextSatelliteId}`);The target is seamless satellite-to-satellite transition. A prediction only tells the network when to switch. The harder requirement is making the switch without dropping sessions, and that is why handover management is listed as a distinct feature of the standard. Beam management is the companion problem. Spot beams must be steered and shaped to serve users unevenly spread over the footprint, and optimizeBeamSteering and wia-comm-005 beam-steer --users 1000 --coverage-area 500 address this.
Constellation geometry. Keeping coverage continuous means distributing satellites evenly. The standard's phasing relations are Phase Angle = 360° × (plane_number / total_planes) and Satellite Spacing = 360° / satellites_per_plane. Its worked example is Starlink Phase 1, with 72 planes of 22 satellites each, 1,584 satellites in total and 5° between planes. The same figures appear in the SDK's SatelliteNetwork constructor (altitude 550 km, inclination 53°, 72 planes, 1,584 satellites). A constellation is therefore described by a handful of parameters, which is what makes it possible to exchange constellation descriptors between systems. Phase 1 of the standard defines canonical envelopes for such descriptors, for inter-satellite link records, and for ground-station and user-terminal data shapes.
Inter-satellite links. Without ISLs, every packet must come down to a gateway near the user. With them, traffic can cross the constellation in space. The standard describes both technologies:
| Property | Optical ISL | RF ISL |
|---|---|---|
| Carrier | 1550 nm | Ka-band 23–27 GHz, V-band 60 GHz, E-band 71–76 GHz |
| Data rate | 1–100 Gbps per link | Ka-band 100 Mbps–1 Gbps; V-band 1–10 Gbps |
| Range | Up to 5,000 km | Not stated |
| Notable trait | Beam divergence under 10 microradians; no frequency licensing; hard to intercept | Wider beams; spectrum must be coordinated |
The optical beam is so narrow that pointing becomes the dominant difficulty. The standard's link budget, P_rx = P_tx + G_tx + G_rx - L_space - L_point, includes an explicit pointing-loss term for this reason. Two spacecraft moving relative to each other must hold the beam on target to within microradians. The mesh that results reduces dependence on ground stations, lowers latency through space routing and improves resilience. For routing, the specification lists OSPF-based approaches, BGP for inter-constellation routing, and DTN (Delay-Tolerant Networking) for deep-space cases. The open challenge is that the topology changes continuously, so routes must be recomputed as links form and break.
Ground segment. The standard calls for distributed gateway and teleport design. A user terminal with a phased-array antenna steers electronically to follow satellites without moving parts, and the modem must keep pace with the handovers described above. Gateways must be positioned so that traffic always has somewhere to land, which is the reverse side of the ISL trade-off.
Regulation and debris. The standard requires compliance with the ITU Radio Regulations, which govern spectrum use and coordination between operators sharing bands. It also requires orbital debris mitigation. With constellations that may exceed 40,000 satellites, debris control is a design requirement from the start. The CLI offers wia-comm-005 debris-check --altitude 550 --inclination 53 to check a proposed orbit.
The difficulties of satellite internet follow the physics of the signal path:
The standard answers these with a shared vocabulary of constellation, link and terminal descriptors, plus calculation tools (link budget, Doppler, handover prediction, beam steering and debris check) available in both the TypeScript SDK and the CLI. A single, auditable formula set matters because operators who build the same constellation but compute the budget differently cannot compare designs or interoperate. With the problems now laid out, the chapters that follow turn to the technical foundation: how the canonical data shapes and interfaces encode this orbital and link information so that independent systems can exchange it.