Vitalik Buterin’s concept of decentralization

This post is a TL;DR of Vita­lik Bute­ri­n’s ana­ly­sis of the meaning of the term «decen­tra­li­za­tion». It is a gre­at heu­ri­stics for ana­ly­sing the archi­tec­tu­re of vario­us sys­tems, inc­lu­ding «social machi­nes», and so I’m leaving here a distil­led ver­sion of the main cate­go­ri­za­tion along with some com­ments rela­ting it direc­tly to the topic of block­cha­in-based «social machines»

Łukasz Jonak, Ana­li­tyk DELab UW

Vita­lik pro­po­ses that the­re are three dimen­sions of decentralization:

  • archi­tec­tu­ral – whe­ther the sys­tem is distri­bu­ted among many computers/servers

  • logi­cal – whe­ther sys­tem logic, data and user inte­rac­tion pat­terns are distri­bu­ted within the phy­si­cal sub­stra­te of the system

  • poli­ti­cal – whe­ther the con­trol over either the phy­si­cal or logi­cal (or both) lay­er of the sys­tem is distri­bu­ted or con­cen­tra­ted in one hand.

One way you can use this cate­go­ri­za­tion to think abo­ut real-life exam­ples is to con­si­der a pro­vi­der of a block­cha­in-as-a-servi­ce, sel­ling a clo­ud-based distri­bu­ted led­ger archi­tec­tu­re to users or con­sor­tia of users. In this case, the archi­tec­tu­ral distri­bu­tion of the sys­tem could be pre­se­rved; even tho­ugh the servi­ce is clo­ud-based, the actu­al com­pu­ters run­ning nodes could be loca­ted in vario­us phy­si­cal data cen­ters. The logi­cal lay­er of the sys­tem would be cen­tra­li­zed (eg. by sta­te sha­ring, con­sen­sus algo­ri­thm), in accor­dan­ce with gene­ral block­cha­in design principles

Poli­ti­cal decen­tra­li­za­tion would be ano­ther mat­ter tho­ugh. The servi­ce pro­vi­der ulti­ma­te­ly con­trols sys­te­m’s infra­struc­tu­re on the phy­si­cal level at least; it can turn it off any­ti­me. Also, unless the client is given the oppor­tu­ni­ty to cryp­to­gra­phi­cal­ly secu­re the­ir assets so that even the pro­vi­der can’t read them, the pro­vi­der would also have some sort of cen­tra­li­zed con­trol over sys­te­m’s con­tent and pro­ces­ses. This is in con­trast to the way public block­cha­ins work: in most cases the­re are no such cen­tra­li­za­tion of poli­ti­cal power over blockchain.

What is inte­re­sting abo­ut Vita­li­k’s cate­go­ri­za­tion is that it goes bey­ond tech­ni­cal lay­er of block­cha­in. It ack­now­led­ges that off-cha­in gover­nan­ce, mani­fe­sting itself among other things thro­ugh the degree of poli­ti­cal cen­tra­li­za­tion, is an impor­tant ele­ment of block­cha­in sys­tems – the­ir tech­ni­cal lay­er is wrap­ped by social lay­er, and only a part of it is regu­la­ted and enco­ded by on-cha­in model of human beha­vio­ur. In social machi­nes, hybrid, human-tech­ni­cal sys­tems, it is not just the tech­ni­cal com­po­nent that con­trols the human; the oppo­si­te also takes pla­ce and the social «wetwa­re» cri­ti­cal­ly influ­en­ces the boun­da­ry con­di­tion of tech­ni­cal component.

Ano­ther inte­re­sting thing is that this con­cep­tu­ali­za­tion of cen­tra­li­za­tion can be used to descri­be pure­ly social sys­tems (check the ori­gi­nal publi­ca­tion for some exam­ples). As a socio­lo­gist, it would be inte­re­sting to find out how com­pa­ti­ble this cate­go­ri­za­tion is with the exi­sting, tra­di­tio­nal ways of thin­king abo­ut dif­fe­rent kinds of socie­ties and poli­ti­cal systems.

Autor pro­jek­tu: Łukasz Jonak

Pro­jekt finan­so­wa­ne ze środ­ków pro­gra­mu „Dia­log” MNiSW

Scroll to Top