A fast optical downlink demonstration does not establish a dependable data service. Customers need evidence about weather coverage, shared outages and the deliveries that remain possible when conditions change.
Optical satellite communication can look compelling in a successful demonstration: a large amount of data crosses the link quickly. A customer buying a continuing ground service has a different question. How much of its required data will arrive in time when the demonstration's favourable conditions are absent?
Weather is central to that assessment, but a percentage labelled availability is not enough. It may describe cloud-free observations, a modelled network, an available receiving station or a completed customer delivery. Comparing offers requires knowing which of those claims the evidence actually supports.
A clear-sky trial can establish that a particular configuration transferred data under the observed conditions. It provides useful evidence about that configuration and the delivery chain exercised during the trial. It cannot, by itself, establish how often comparable conditions will occur throughout the customer's service period.
The distinction matters when a supplier presents peak rate and annual availability close together. Readers may combine them into an implied annual data volume even if the numbers come from different configurations, locations or periods. The commercial review should ask whether the two claims were evaluated together and whether both apply to the service being offered.
This is not a reason to dismiss demonstrations. It is a reason to use each evidence type for the question it can answer. A technical trial helps establish feasibility. A representative operating record helps establish repeatability. A service assessment adds the customer's workload, deadlines and purchased access.
In their 2015 research on optical ground-station networks, DLR researchers Christian Fuchs and Florian Moll evaluated five years of cloud observations and considered correlation between sites. They also examined the time behaviour of simulated outages. The study demonstrates why individual-site conditions and their relationship matter; its numerical results do not establish the availability of any current commercial network.
A network map therefore needs an accompanying service explanation. Additional stations can offer useful diversity, but their value depends on whether their unavailable periods overlap and whether the customer can actually use them. Counting locations alone does not measure the extra service opportunity.
The buyer should ask for evidence of the combined offer rather than adding individual station percentages. It also needs to understand whether the evidence describes all advertised sites, only an evaluated subset or a future expansion. A site planned for later inclusion should remain a future dependency in the commercial assessment.
A simpler question often exposes the gap: when the usual receiving location was unavailable in the evaluation, what happened to the customer's data? A useful answer identifies whether it was delivered elsewhere, delayed, partially delivered or unavailable within the relevant period.
Availability measured over every hour of the year may be a poor description of a service purchased for specific data deliveries. A station can have favourable conditions while the customer's service is not scheduled, or be unavailable during the small number of periods that matter to a time-sensitive product.
Consider an illustrative company offering a weekly environmental report. It can tolerate a delay in individual acquisitions if the full dataset arrives before the weekly processing deadline. Another customer needs each daily report before a fixed publication time. The same receiving service could support the first business well while producing frequent missed deadlines for the second.
Their acceptance measures should reflect that difference. The weekly customer might track complete reporting packages delivered before the cutoff. The daily customer might track each required delivery and the duration of lateness. Neither should assume that a general cloud-free percentage directly measures its product availability.
BDI's guide to ground-service handover explains the related delivery boundary. Weather assessment becomes commercially useful when the evidence continues past reception and shows whether the agreed information product reached the customer's destination.
Two services can lose the same total amount of time while creating different business consequences. Several short interruptions may be absorbed by a flexible workflow. A single extended interruption can leave a reporting cycle incomplete or create a backlog that persists after conditions improve.
The supplier should therefore present interruption duration and recovery behaviour alongside an aggregate percentage. A monthly total hides whether missed service was concentrated in a season, repeated at relevant delivery times or followed by prompt clearance of delayed data.
The review should also preserve missing observations in the evaluation record. If weather or service data was not collected during an interval, the interval should not silently become a successful one. Explain how unknown periods affect the reported result and whether they coincide with conditions the evaluation most needed to observe.
A record gathered during a favourable season can still be useful as preliminary evidence. It should be described that way. Extending the evaluation across materially different conditions adds information that cannot be supplied by increasing the precision of an annualised estimate drawn from a short trial.
ESA's April 2026 account of the VISTA project describes sky-monitoring technology evaluated in an optical-ground-station setting, including work at DLR's Trauen station. The account also identifies further development needs involving seasonal and climate coverage and forecast refinement. It supports the existence and direction of that development work, rather than a guaranteed commercial availability level.
A customer evaluating a forecasting feature should ask what decision it improves and how that improvement was measured. Producing a forecast is different from demonstrating that the service delivered more of the customer's required data. An accurate prediction of an interruption may help the operator plan, but it does not remove the interruption.
Useful evidence compares the service outcome with and without the feature under a documented evaluation. It should retain cases where the forecast was uncertain or wrong, and describe how those cases affected delivery. Otherwise a visually convincing dashboard can become a substitute for evidence about the commercial result.
An optical network may include facilities operated by different organisations. A customer should establish which portion of that network is available under its proposed agreement, what scheduling or capacity limits apply and how the provider reports a service that cannot be accommodated. An impressive infrastructure footprint is not automatically a reservation.
The same distinction applies to alternatives during poor conditions. A proposal may include a different delivery path, later reception or another service tier. Each alternative should explain the changed outcome and any additional cost. Deferred data delivery can be a suitable solution for an archive product while being inadequate for a deadline-driven subscription.
The SSC Copernicus ground-station service extension illustrates the broader market importance of dependable ground infrastructure. Its programme context does not establish the terms or weather performance of a separate optical service. Those require evidence from the actual offer.
The charging unit deserves the same review. If the customer pays for reserved opportunities, compare expenditure with accepted deliveries, including occasions lost to weather. If it pays for delivered volume, establish whether late or incomplete data is counted. The price is meaningful only beside the outcome that triggers the charge.
A defensible comparison brings the pieces together: the configuration demonstrated, the conditions evaluated, the access purchased and the customer deliveries completed. Peak speed remains useful, but it becomes a commercial advantage only when the service can turn available opportunities into the information the customer needs within the period that gives it value.
DLR research is used for the relationship between cloud observations, correlation and simulated availability, not as a prediction for a current network. ESA's VISTA account describes a development project and remaining validation needs; commercial examples are original analysis.
Suggest a correction