Approaching the issue of fairness requires a process of definition in which each stage could lead to a different outcome. As with most issues in philosophy, there is no answer, so I will attempt to outline the defining process.
Firstly is the question of whether fairness is judged by the outcome of the distribution or by the rights of the individuals among whom the consequences are distributed. Consequentialism holds that it is the former, Virtue Ethics the latter. Complicating the matter Consequentialist philosophers like Mill consider some virtues necessary to issues of fairness and those following Kant consider that some form of de-culturalisation of virtue is necessary by generalization (roughly, act only in a way such that if everybody acted that way you world be happy in such a world).
Taking the view that what is fair is based on the consequences of the distribution, one must then decide the consequences on whom. Strict Utilitarianism holds that fairness would be that which gives such share as to maximise the benefit to humanity as a whole. This can be contrasted with Ethical Egoism which holds that the consequences of a distribution on the agent of that distribution is the only rational consideration, rendering fairness as it is commonly conceived redundant as a consideration. Positions obviously exist along the range between these two extremes. The problem with deciding fairness under any strict consequentialist philosophy is that it is dependant on the agent's knowledge of the consequences which in all cases is limited to a greater or lesser extent.
Virtue Ethics avoids the problem of limited knowledge of consequences by pre-supposing that there is some guide to what is fair. This can take the form of established virtues as facts (an old Aritaic idea), use of our own feeling of fairness to judge how others probably feel (from Kant) to various versions of living according to pre-established rules on which all members of society agree, but which everyone accepts may not be fair in any given exceptional circumstance (such as Rule Utilitarianism).
Describing what fairness is, rather than what it should be, modern experiments in psychology have shown that humans have an intrinsic feeling that fairness is important (in terms of equal treatment) even to the detriment of their current position. This is extended even to chimpanzees as Franz De Waal has shown. The indication is that we, as a species, are not consequentialists in a rational sense, but consider unfairness to be unacceptable and reprehensible instinctively even at cost to our own immediate reward. This can be reconciled with consequentialism by for example, considering our instincts as guides developed by evolution to that behaviour which yields the best consequences for our survival.
In short, fairness seems to be a hard-wired concept among social species and so must serve some purpose which promotes our individual survival. Thus fairness can be reliably evaluated by our instinct in a way which should benefit at least our social group (if not society as a whole), but such evaluation is rarely possible in complex cases where knowledge is incomplete, and reliance on cultural rules or virtues becomes necessary.
