A temporary solution can be reasonable during an emergency and become an unjustified infrastructure as time passes. That change rarely takes place in a visible ceremony. It happens when a contract is renewed through inertia, an account remains active although nobody reviews it, a data centre expands because available capacity encourages further use, or an unrepairable device stays in service because replacement appears easier. The cloud still looks weightless on a screen, yet it rests on buildings, electricity networks, water, minerals, equipment, and labour. Responsible policy examines the calendar and that material foundation together. It asks how long the exception will last, who verifies that it remains necessary, and how it will be dismantled without transferring the cost to people who have the least power over the decision.
Temporary measures also need design#
An emergency requires action with incomplete information. A school may enable a temporary platform to maintain teaching; a local authority may purchase additional capacity during a crisis; a company may distribute equipment to preserve an essential service. The initial decision is not necessarily mistaken. The problem begins when speed is confused with indefinite permission. If a measure has no review date, accountable owner, exit budget, or closure conditions, time works in favour of permanence. Every renewal adds integrations, habits, and dependencies that make the measure more expensive to question.
An emergency requires action with incomplete information. A school may enable a temporary platform to maintain teaching; a local authority may purchase additional capacity during a crisis; a company may distribute equipment to preserve an essential service.
A well-governed exception contains its own expiry mechanism from the first day. Documentation identifies the specific problem, expected outcome, resources used, and harm that would require the measure to stop. It also distinguishes an expiry date from an evaluation date. The first prevents automatic continuation; the second creates an opportunity to assess evidence before deciding. If the organisation wants an extension, it must show that the need persists and compare renewal with current alternatives, rather than with an emergency scenario that has already ended.
Duration also changes the ethical character of a decision. Limited data collection may be proportionate for several weeks and excessive after several years. Loaned equipment may answer an immediate need but lock a community into proprietary parts as it ages. Professional analysis therefore asks more than whether a tool works. It examines how long it works, for whom, with what cumulative consumption, and under what practical possibility of withdrawal.
The cloud has a calendar, a territory, and a material base#
Digital services are commonly represented as instantaneous flows, although they depend on assets with long life cycles. Data centres require electricity, cooling, land, connections, and server replacement. User devices require raw-material extraction, manufacturing, transport, maintenance, and treatment at the end of use. An efficiency improvement per operation therefore does not guarantee an absolute reduction in impact. If the cost of each task falls while the number of tasks multiplies, total demand can rise. The European Environment Agency specifically notes that digitalisation can deliver environmental benefits while also creating pressures and rebound effects that require governance.
The International Energy Agency adds another scale to the discussion through its analysis of data-centre electricity consumption and the geographical concentration of its effects. A global figure may appear moderate while concealing local pressure on an electricity network, water, or land. Responsible decisions consequently need information broken down by location, period, and type of workload. They also need to acknowledge uncertainty: projections depend on technology adoption, hardware efficiency, service models, and energy choices.
The material dimension continues after an application is switched off. The Global E-waste Monitor produced by UNITAR and the ITU identifies a persistent gap between electronic waste generation and documented recycling. That finding makes repairability a matter of social infrastructure rather than a nostalgic preference. Accessible manuals, available parts, disassemblable design, long support periods, and repair markets preserve value, reduce premature replacement, and distribute capabilities beyond the original manufacturer.
Expiry, review, and the authority to close#
An expiry clause works only when somebody has genuine authority to enforce it. The accountable person cannot merely send a report to the same department that is rewarded for expanding the system. That person needs access to full costs, usage data, incidents, contracts, energy consumption, and exit conditions. The role must also be able to suspend a renewal while concerns are resolved. Independence does not require isolation, but it does require sufficient separation among the party that proposes, the party that sells, the party that operates, and the party that decides whether the measure continues.
A review combines quantitative indicators with situated experience. Relevant indicators include total consumption, equipment life, repair rates, the volume of retained data, licence costs, support hours, and the number of people who leave the process. Situated experience explains what an average erases: who needs a better connection, who absorbs unpaid tasks, which neighbourhood hosts the infrastructure, or which repair shop is excluded by unavailable parts. Participation is useful when it can alter the decision, not when it merely validates a conclusion already made.
Closure requires preparation. Data must be exportable in usable formats; essential configurations need documentation; equipment must be reassigned, refurbished, or recycled through traceable channels; and affected people need to know the timetable and alternatives. An exit plan funded from the outset prevents dependency from later being presented as an unavoidable argument for continuation.
Case 1: a school platform born in an emergency#
A composite case illustrates the problem without attributing it to a particular institution. A school adopts a platform within a few weeks to preserve communication during an exceptional disruption. The tool answers a real need. Two academic years later, circumstances have changed, but the platform retains every function, collects more data than necessary, and requires recent devices from some families. Renewal is processed as recurring expenditure, so nobody compares the service again with simpler channels.
The review identifies three costs that emerged over time. The first is financial: licence numbers have grown alongside inactive accounts. The second is material: several functioning devices cannot run newer versions. The third is distributive: families with limited connectivity spend more time completing routine procedures. The school does not need to deny the tool's initial value in order to change direction. It can retain only justified functions, reduce retention, guarantee an equivalent in-person channel, and extend device life. An annual decision date, with students, families, teachers, and technical support represented, changes automatic renewal into an explainable choice.
Case 2: computing capacity that becomes permanent demand#
Another composite case concerns a public administration that purchases exceptional computing capacity for a limited public campaign. When the campaign ends, the reserve remains available and several departments create new processes to use it. Each project looks small; together they consolidate a level of consumption that was never authorised as permanent policy. The dashboard reports a falling cost per operation, but it does not highlight growth in total consumption or the location of the infrastructure.
A mature evaluation reconstructs the baseline and assigns every workload to a current purpose. Redundant processes are withdrawn, flexible tasks move to periods that are suitable for the grid, and essential services receive explicit priorities. The administration requests comparable information about energy, water, equipment, and emissions without turning a single metric into a sustainability certificate. It also publishes who may approve additional capacity and which threshold triggers another assessment. Efficiency retains its value, but it no longer serves as an excuse for growth without deliberation.
Case 3: temporary sensors that nobody knows how to remove#
The third composite case takes place in a municipality that installs environmental sensors during construction work. The measure was intended to last eighteen months and provide evidence about noise and mobility. The works end, but the devices remain active because different teams discover secondary uses. The connectivity contract is renewed, data accumulates, and some sensors fail without a clear inventory. Residual usefulness exists, although it no longer matches the original mandate.
The municipality organises a public review by location and purpose. Sensors required for a clearly defined environmental series enter a stable programme with maintenance, data minimisation, and a budget. The remaining units are removed; repairable equipment moves to other projects and the rest enters a documented treatment channel. The decision includes an accessible register of devices, accountable owners, and forthcoming review dates. The case demonstrates that partial closure is not a failure. It is a way to retain what is useful without turning every technical possibility into a permanent obligation.
Measuring a full life and sharing the capacity to decide#
The central metric should not be novelty or isolated performance, but the service delivered over a complete life. That perspective brings together acquisition cost, energy, maintenance, repairs, updates, human time, replacements, and end-of-use treatment. It also prevents an organisation from claiming a saving after shifting work to families, small suppliers, or support staff. Honest accounting records visible resources as well as displaced burdens.
Equity requires comparison of impacts across groups and territories. A solution can save time for the average person and create a severe barrier for a minority; it can reduce operational emissions at one office while increasing device turnover in hundreds of homes. Disaggregated data needs protection so that it does not expose individuals, but that precaution does not justify erasing differences. Surveys, interviews, complaints, and accessibility testing complement technical records.
Freedom depends on the presence of an equivalent alternative. A route that requires personal connections, specialised knowledge, or indefinite waiting is not a genuine exit. Participation in turn requires understandable information before a decision, time to respond, and feedback explaining what changed. When a community can propose limits, request evidence, and trigger a review, sustainability stops being a supplier's label and becomes a shared practice.
Repair before replacement, learn before expansion#
European repair policy reinforces an important direction: extending the use of repairable goods and reducing their premature disposal. Within an organisation, that direction translates into purchasing criteria that value parts availability, documentation, warranties, security updates, disassembly, and reasonable repair costs. Initial price ceases to be the only criterion. A slightly more expensive device may be the better choice when it remains useful, accepts local maintenance, and retains value at the end of a contract.
Repair also describes an institutional response. If a temporary measure has produced exclusion, unnecessary spending, or data accumulation, the solution does not end when it is switched off. The organisation must correct records, restore options, compensate burdens where appropriate, and amend the procedure that enabled automatic continuation. A closure report can document what was removed, what was retained, which resources were recovered, and which signals will be monitored in the future.
The decisive question is not whether technology deserves enthusiasm or rejection. It is whether the organisation retains the ability to understand its footprint, limit its duration, and change its mind. A temporary solution preserves legitimacy when it includes a date, evidence, an accountable owner, an alternative, and repair. Under those conditions, time does not silently transform an exception into destiny; it provides information for a fresh decision.
