Why a Distributed System for the Trading Industry instead of a Centralised One
In our previous posts, we talk about the inherent inefficiencies with the Commodities Trading industry based on paper receipts, and how our project on block-chain based trading system could solve this through the use of smart contracts on Ethereum blockchain.
In this article we go about why other existing digital solutions based on centralised systems are not very effective compared to distributed systems for receipt and trust management in the commodities trading industry.
Governance
Who should manage the digital system? It needs to be a third party trusted by all organisations involved (traders, warehouses, shipping companies, banks, governments, etc). This entity has to be trusted to not misuse its power or cheat other involved parties (e.g. charge high fees for consumer access). This central node will intrinsically have the greatest power as denoted by graph theory and market forces model. This makes the market inefficient and reduces competition which is bad for the consumers of this technology.
It is difficult to find such central entities trusted by all parties, the few that qualify tend to be monopolists with great market power. Governments might immediately come to find but they are inherently efficient because of the lengthy bureaucratic procedures involved.
Multi-version concurrency control
Centralised systems only have a single version control lacking decentralised multi-version concurrency control mechanisms that maintain consensus about the existence and status of shared facts like equipment, loan, payment receipts and contracts. In central systems, all of the verification has to be done by the central node and if the central controlling entity decides to cheat others, there are no other versions to crosscheck these claims especially in the Trading industry.
Distributed systems allow for a secure and encrypted trade and receipt verification system be it for use as collateral for leveraging in banks or to satisfy payments in transactions.
Untrusted Markets
Central databases are not good when operating in environment with fraud issues where parties do not trust each other. Trading markets usually have a high probability of receipts fraud. And, participants in these transactions, the sellers and the buyers intrinsically do not trust each other. Especially in the case of warehousing and shipping receipt tampering, bank loan collateral proof fraud.
Single Point of Failure
There is a lot of pressure and responsibility on the central governing authority for the centralised database system. Like any centralised system it has the vulnerability of a central point of failure which is also easy for hackers to target.
If the one central system fails, the entire network collapses.
For example: The Amazon Web Services collapse in April of 2011 that took down Reddit, Quora, and NYT among 70 other sites. This crash in Amazon’s Northern Virginia data centre brought down the entire AWS based network in the Northern US. These outages theoretically cost Amazon $66,240 per minute — or nearly $2 million — based on Amazon’s net sales.
Local control
Just like the government in a country, central systems lack local control. Distributed local systems relieve stress and responsibility on the central node. But distributed systems allow huge portions of data and control to be handed to the local sites (servers) and to the local owners of the data. This will be very important in commodities markets in developing nations where the governments or central authorities are inefficient in maintaining full widespread control and efficient governance.
Modular growth
Central database systems do not offer resilient modular growth. The central point has to be interrupted to add new sites or remove sites. While for distributed systems, addition and removal only concerns local neighbouring sites allowing for increased level of growth to accommodate emerging and expanding trading markets especially like the trading markets.
Expensive local communications
Local communication costs are costlier in resources and time (Think of bureaucratic delay in passing laws and reforms in the government ). All information has to pass through a central authority.
Low Latency
Latency and response times for the electronic receipt verifications are much lower In distributed systems like the one implemented in our project.
References:
Iacob, Nicoleta Magdalena. Moise, Mirela Liliana. Case study Paper: Centralised vs Distributed Databases. Faculty of Finance, Banking and Accountancy, Dimitrie Cantemir Christian University.
Easley, David. Kleinberg, Jon. Networks. Crowds, and Markets: Reasoning about a Highly Connected World, Dept. of Economics Cornell University.
