Snowden, Dilma Rousseff’s mobile, Belgian Foreign Affairs, Bundestag, European Commission, Elysée: it is hardly a month without a "case" making the headlines. The scenario is unchanged: such state institutions have been spied for years - we do not know the authors - X is suspected (your choice: United States, Britain, Russia, and China). The only variation is due to the reaction of the victim: wrath, state affair, or embarrassed silence.

Most surprising, however, is that "allies" are spying on them. But it is logical as the conditions of cyberspace invalidate traditional alliances, for two main reasons. First, in a traditional alliance, whatever the political and strategic objectives, it is first necessary to add up the forces: troops and armaments are highly tangible, visible, countable, and assessable. However, the power of cyber is not counted in alignments of bytes or computers. It is based primarily on the creation of highly skilled teams, and that ultimately needs little equipment to work and progress. In other words, it is very difficult to add up intellectual capacity.

Especially, a second element is coming up: despite the appearances of publicity, opening and voyeurism of the Internet, facing the average user of cyberspace, it is a hidden, opaque, discreet space. It is very easy to act anonymously in cyberspace: not only to not be detected but even to impersonate others. Also, in all cases of recent years, we never had absolute technical proof of the alleged perpetrator’s responsibility.

This is an unprecedented strategic novelty. In the world we were used to, we knew who the enemy was, so who was a friend. Admittedly, the criteria could be imperfect, alliances could vary in time, and unlikely compromises could become possible. In cyberspace, you never actually know who is actually acting. The actor is always an unknown. Therefore, one cannot certainly describe it as an enemy. But if one cannot designate the enemy, then it is equally difficult to identify the friend, then the ally. This gap identification affects the whole mechanics of alliances.

This does not mean that there are no alliances in cyber. Simply, they are hidden, discrete, usually bilateral, confined to strictly defined and limited objectives. When two on a given project, we necessarily know who is who: oneself, the ally, the others. The mutual identification procedures can reveal "the others".

So, we can have a system of bilateral alliances. Me, country A, I ally myself with X on such a topic, with Y on another. But suddenly, my cooperation with X can affect Y with whom I have also another project.

Thus the embarrassment of many official reactions can be understood, as victims of a power with which they have also co-operation programs. Now, in the simplistic media world, this ambiguity is difficult to explain, even to justify. Thus, while one trumpets the myth of general alliances, one whispers that there are no allies in cyber, and restricted alliances are practiced in the greatest secrecy.

Olivier Kempf published Alliances et mésalliances dans le cyberespace (Economica, 2014), awarded the Grand prix du livre cyberstratégie 2015.