To decentralise or not decentralise, that is the question …


I must apologies to William Shakespeare for abusing his famous quote. However, this post will start with the following like every good fairy tale.

Once upon a time, a network of networks became interconnected using a suite of protocols referred as TCP/IP. Named the Internet, this network of networks started to decentralise data across major cities of the United States. So, should a nuclear bomb would destroy computing facilities the data would not be lost. The army originally relied on it, to prevent data loss for a threat that thankfully never occurred. However, members of academia thought it would a good idea to connect to this interconnected network of networks. Information and data can be shared more easily and community of researchers can community can communicate with each other more easily. That cannot be bad and it remains a great idea.

In this last paragraph, three concepts keep being mentioned; i.e., network, communication, and data sharing. We need to remember networks relies on telecommunications tools. For example, when Cambridge University in the United Kingdom, needed to interconnect all their colleges on the same network. Cambridge streets and pavements were dung up to bring the cables to connect these buildings. The latter are scattered all around the city centre. When, the Internet needed to be connected to the other continents, then some cables were layer under the oceans. This map was obtained from submarine networks and shows a recent maps of the Internet under the oceans. It looks amazing and now some network of satellites is growing too…..

All these telecommunications cables, satellites and other specialised equipment, such as routers and TCP/IP would not be useful without a real purpose. So both ideas goes hand-in-hand; the need of data sharing is removed, and the need of any forms of networking becomes irrelevant. The more networks developed, the more data sharing becomes even more relevant.

Data and information sharing is not a trivial task. A bit like a postal services, communication takes place between at least two or more parties. Information is exchanged using a language that is understood by all parties. Well the World Wide Web is based on these ideas. If you think of it, we have some addresses and name – i.e., IP addresses indicating where to find a web server and a name attached to a web server. So, our web servers store information and send them back to the clients (http). We use some languages or protocol to requests and respond to these requests. We some formatting and structure languages to make this information more human understandable. Some other more complex techniques encode multimedia content, such as this post. At that state, we are using some software elements.

Remember without the Internet, the World Wide Web and other peer-to-peer systems would not exist. We have decentralised information across some web servers, that are accessed using the Internet. Before the World Wide Web, other applications used the Internet. Our email systems relies on a different set of protocols lying in the application layer of the TCP/IP stack; i.e., SMTP, POP3, and IMAP. FTP has been used for years to share files and data. Secure Shell has been useful to connect remotely to systems too. For a most extensive list, this website should help and you will see something referred as Bitcoin. Satoshi Nakamoto describes Bitcoin as a peer-to-peer electronic cash system. If you read the original paper, you will find some ideas that may have been used in communications and networking. Some distributions and decentralisations occurs through bitcoin servers, using the Internet..

Bitcoin the paper: https://web.archive.org/web/20140320135003/https://bitcoin.org/bitcoin.pdf

Data sharing can often be related with encryption. Through history, any confidential data, messages and ledgers may have been encrypted and encoded to prevent any eaves-dropper understanding its content. So, the idea of crypto currencies is likely to rely on some forms of encryption too. The idea of blockchain originated in 1982 and the quote below should introduce well its design.

Block chain idea:

A distributed computer system that can be established, maintained, and trusted by mutually suspicious groups. It is a public record-keeping system with group membership consistency and private transaction computations that protects individual privacy through physical security. The building blocks of this system include physically-secure “vaults,” existing cryptographic primitives (symmetric and asymmetric encryption, cryptographic hash functions, digital signatures), and a new primitive introduced by Chaum—threshold secret sharing [Cha79].

Blockchain: https://arxiv.org/pdf/1810.06130.pdf

On the same idea, IPFS, Elastos, and Solid decentralises into nodes some elements. What is interesting some of the marketing content from Elastos confuses the main ideas by misusing some technical terminologies. It is suggesting a new Internet became in existence. However, it is a networked operating systems that relies on the Internet itself. It perceives the telecommunication network perceived as old-fashioned, but cannot work without it. Solid and IPFS are more interesting as they both aim to keep data safe by distributing the data into wallets or pods. Instead of centralising data within a service, such as Facebook, Netflix, or other websites, we all become data governers of our data. Netflix and Facebook will need to get to our pod to gain access to the data. Us users can become more in charge of our data.

So far we have discussed about decentralisation. The latter occurred in routers and routing routes, domain names servers, web servers, distributions of data through nodes, wallets and pods. However, some form of centralisation has occurred with the development of Cloud technology relying a data centers. Instead of organisations working with their own IT facilities, they use computing facilities accessible using the Internet and manageable using a web interface. A centralisation of computing facilities in large data centres has become the backbone and reliance for our online shopping, entertainment and communication (video calls). Data, software, and services is now stored and made available through some humungous computing facilities, sharing hardware through a form of virtualisation. Some healthcare Cloud systems have now been offered by the major big companies too. The world is now becoming ever more centralised on the Internet and large data centres. It is a very scary thought of the powers it can bring in the wrong hands. I must have watched too many James Bond films….

To conclude, it is worth to mention an effort to bring the computation to the data is taking place. No consensus exists yet. However, computations needs adapting to prevent inferential discovery of sensitive data. Data preservation and protection becomes even more important. This unpublished paper illustrates some aspects of these ideas.

,

Leave a comment