One of the worst nightmares for the best blockchains, based on decentralized and distributed systems, is to become vulnerable because of a “byzantine fault”, i.e. “a fault presenting different symptoms to different observers” (Ayala 2016, p. 27)1.
To counter this in fully automated systems, some companies developed a strategy named “Byzantine fault tolerance”, i.e. services assuming there are not too many components being in a faulty state. As an example. Ayala quotes “some aircraft systems, such as the Boeing 777 Aircraft Information Management System and the Boeing 787 flight control systems, use Byzantine fault tolerance”. The problem is nevertheless persistent when humans have a word to say, and may lead to a “Byzantine failure hack”, meaning that when facing problems, the control center (human or automated) fails to reach an agreement with the components (human or automated) building the system itself, resulting in a dramatic failure of the whole system and creating a door wide open to hackers.
But what is the origin of this expression? It is in fact a shortcut of a longer tile, invented by three top researchers of SRI International. The final result of their study, commanded by the US Army Research Office on ballistic missiles systems, was published in 1982 as “The Byzantine Generals Problem”2, synthetized in the four following schemes:
But is it true that historically this was a real problem for Byzantine commanders? Not at all. The authors, probably did not have enough insight of the Byzantine Empire’s army. In fact, the empowered generals had full power. The best of them were among the most brilliant the world ever knew, such as Flavios Belisarios (505-565), who alone, doubled the surface of the Empire, taking possession of more than half of the former Roman Empire, thanks to a rock-solid and perfectly commanded army. The Empire, born in 330 as the Eastern Roman Empire, proved a great one, with great generals, at least until 1204.
Yet the brilliant IT researchers were not totally wrong on their choice of title. The problem they underlined and alluded to appeared only after 1203, when the Crusaders from acrosl Europe decided not to fight anymore for Jerusalem and to stop at Constantinople.
After the felony and the complete sack of the capital, Constantinople in 1204, itself by those Christians – the famous San Marco bronze horses stolen from the hippodrome of Constantinople were then taken to Venice. – After this point, Byzantium and its Empire never really recovered economically and military, even if its final falletook place 250 years later, in 1453, when Constantinopleewas taken by the Ottomans.
During those last 250 years, Byzantium had the strongest army of all Christianity, but could not go alone to war against its different enemies. Byzantine generals commanded in fact only their troops, the rest being constituted of different Christian groups each led by its own warlord and not a subject of the Byzantine general. This led to many Treasons, misunderstandings, huge mistakes, which further led to many victories but at the cost of a massive lost of men for the Byzantine army and also many defeats which turned to a complete disaster, due to the early hurried attacks or unannounced retreats by those auxiliarys groups.
It is this fact the IT specialists pointed out. The core problem beingdthe very sad lack of cohesion of the main allied commanders, with the byzantine generals, oveg two centuries. For those interested, this phenomenon is masterly depicted by S. Kyriakidis3, one of the most prominent scholars of this period.
Coming back to the title and therefore the IT problem, we can hence observe the simple lack of an adjective: late. the correct title would have been The late Byzantine Empire Generals Problem.
Author: Laurent Chrzanovsk