Centralized vs. Decentralized: The Best (and worst) of Both Worlds
Since its inception in 2008, blockchain continues to be a source of groundbreaking solutions and possibilities. The concept of a decentralized system (i.e. Bitcoin) that successfully implemented a digital peer-to-peer system was a much-needed breakthrough, especially after decades of domination and exploitation by increasingly irresponsible and centralized financial giants. The merits and demerits of blockchain technology have been discussed at great length, with much debate over how feasible decentralization truly is.
Often, decentralized systems are confused with distributed systems. A distributed system is a model in which components located on networked computers communicate and coordinate their actions by passing messages. With this definition, a decentralized system is also a distributed system. A key point about decentralization, Blockchain enthusiasts insist, is that there’s no central point of control.
To understand what decentralization truly entails, here is an interpretation of here’s Vitalik’s (2018) take on the three aspects of decentralization, that are theoretically independent and hence have independent consequences from a practical standpoint:
1. Architectural decentralization: How many physically independent hardware systems exist in the network? What is the number of the physical computers present in the network? How many of them can breakdown at the same time without jeopardizing the entire system? This aspect of the decentralization is what determines how distributed a network is.
2. Political decentralization: How many decision makers does the network function with? How many individuals or parties ultimately control the computers that are part of the system? This aspect of decentralization determines the trust aspect of the network i.e. how many parties are being trusted by the network.
3. Logical decentralization: On how many different points are the network resources focusing on? Is there consensus in the network about the purpose of the network? Can the data structure and interface, function the same way independently, if the system is split into two parts? This aspect of decentralization deals with “consensus” in the network i.e. what the network agrees on, or for that matter disagrees on. Logical centralization means that the network agrees on one thing, or has one truth for example, the network has consensus on how much currency you hold in your wallet. Logically decentralized implies that the network can have multiple variations of values for 1 thing, i.e. carrying on with the same example, the network does not have consensus on how much currency you hold in your wallet. Although apparently useless in utilitarian terms, logical decentralization is quite useful in certain limited contexts, none of which are relevant for this discussion though
You will notice that theoretically, blockchain technology in this respect enjoys architectural and political decentralization, i.e. multiple computers in multiple locations worldwide (architectural decentralization) and no single party controls the software running on the network (political decentralization). It is, however, logically centralized because the entire system works like one, and although decentralization purists continue to push for logical decentralization, logical centralization ensures that the entire system is working towards a common goal.
But why does this all even matter? There are three fundamental arguments in favor of decentralization and blockchain:
1. Fault Tolerance: Because they rely on many separate components, decentralized systems are less likely to fail accidentally.
2. Attack resistance: Due to the presence of a lot of players, decentralized systems lack central points of failure; there’s no one point of attack that would disarm the entire system. This makes it more expensive and less viable to destroy these systems.
3. Collusion resistance: In contrast to centralized systems, it is highly unlikely for participants in a decentralized system to be able to collude for selfish goals. This is unlike large corporations being able to lobby to push forward their own agenda.
Each of these properties scales with the level of decentralization in the network. The more decentralized the network, the more fault tolerant, attack resistant, and collusion resistant it becomes.
The hype around the concept of decentralization is not baseless. Decentralization facilitates diversification of activities, where sub autonomy to constituents can ensure innovation and speedy success tactics. Under an overall coordination with the upper management, these sub autonomous constituents can quickly become self-sufficient and profitable.
Decentralization also encourages the evolution of employees to capable managers and leaders by boosting personal responsibility and accountability. This is a solution to the challenge many companies face of a dearth of managerial talent, which acts as a limiting factor to their growth. An organization cannot expand effectively beyond the scope and abilities of its managerial personnel.
Now consider the issue of network governance or community management. Blockchain or other politically decentralized networks use the community as a deciding body, and this ensures safe play and transparency. This type of implementation is in contrast to centralized networks where numerous players are involved in a form of hierarchical decision-making and the filtering and ranking of data, where centralized unaccountable parties hold unequivocal power to change the terms and conditions of the game at their volition.
As opposed to a decentralized system, a centralized one can become a potential site of attack or become an attacker itself. It is easy for players to collude to benefit themselves at the expense of others in centralized systems. Additionally, a centralized system propagates a mono-culture of information, as the central node starts to control what information is trusted or ranked higher than the other.
However all of this advantage comes at a cost. The three major drawbacks of decentralization are:
1. Lack of focus: Too much emphasis on independent decision-making at each level of the system can potentially create ambiguity about principal objectives and reduce their importance. Individual decision-makers may take actions that benefit their segments and not the system as a whole. In centralized systems, central bodies make strategic decisions and the rest of the system follows; in decentralized systems, governance is often complex and decision making is slow and drawn out.
2. Duplication of work: Decentralized systems are secure by nature through the concept of redundancy. Each node (participant) in the system repeats the same task that all of the others do. This creates an uneconomical system both in terms of monetary cost and energy.
3. Speed of action: Another tradeoff for the security and innovation that come with decentralization is the expected loss of speed. This exchange is prevalent in many other areas of life. For example, if you go out to eat alone, you often find it easier to decide what to order; however, if you go out with friends, attempting to reach some level of consensus takes a long time. This also correlates to the first drawback, the lack of consensus, or rather the effort required to reach it, not only slows down the decision making process but removes focus from essential objectives as well.
Given the nature of these trade-offs, total decentralization doesn’t seem to be the most ideal solution for all scenarios, so it might be worthwhile to observe the merits on the other end of the spectrum as well:
It is widely documented that a centralized authority at the heart of a system will optimize the system, whether information is concentrated or dispersed. To fully utilize the underlying information structure in networks in a scenario where information is concentrated, authority over the core node will improve coordination within the system. On the other hand, in case of highly dispersive information, direct control over the central node will ease communication and improve turnover time.
This brings us to coordination and communication efficiency in centralized systems, as the same information is not stored or used multiple times as different blocks. Therefore, partially centralized networks with lack of overlapping information are more efficient and save the costs of generation and maintenance of a giant network system.
Another key advantage worth mentioning of having a central node is that it ensures uniformity and eases coordination on many fronts. It prevents inconsistencies for processes that do not need to be creative with their procedure every time.
Centralization also cuts down on the administrative costs of decentralized systems that exist due to the requirement of trained personnel at lower levels to accept authority and take part in decision making. Moreover, the services of such highly paid manpower is not always utilized fully, particularly in small organizations, and thus leads to a waste of resources and no return on investment.
There is a constant debate whether or not a company should go for (political) decentralization. Companies often grapple with the idea of delegating a certain level of decision-making powers to lower units and keeping the core decisions in hands of the top management. This is an understandably smarter approach when it suits the corporate structure. Multinational corporations, that successfully implement this model achieve regional success due to their local responsiveness. However, trading companies may prefer a centralized system to keep uniformity and maintain stringent checks on their operations for obvious reasons.
However, a problem with decentralization and flat organizational structures is that they cannot be implemented in their extreme form to large companies, as employees quickly become demoralized by the lack of a clear objectives and guidance that comes with doing away with the titles and hierarchy. Zappos adopted this structure aiming to empower its employees, but the company faced the problem of high employee turnover owing to the resulting uncertainty and lack of career growth. Unilever, in contrast, has fared better by keeping its titles and headquarters in place, while creating self-sustaining, semi-autonomous units throughout the globe.
With all the research and experimentation that is currently going into understanding and implementing decentralized systems in the crypto-sphere, there is a lot to look forward to in the near future. It is usually in these times of uncertainty and experimentation that groundbreaking ideas and innovative solutions to commonplace and accepted problems come about.
While decentralization might not be the solution to every problem on earth, it still has far reaching implications for the way we conduct ourselves in society and how we reach consensus in an increasingly digital world. An ideal future would be one where we can combine the speed, efficiency and focus of centralized systems, with the security, incorruptibility, and inherent creative capacity of decentralized ones.