2024.09.17

Toward the Era of Decentralized Web

Masaki Kohana
Associate Professor, Faculty of Global Informatics, Chuo University
Area of Specialization: Web Systems

History of the World Wide Web

More than 20 years have passed since the World Wide Web (WWW) was born, and it has now become an indispensable part of work and daily life. The WWW is a system that allows documents on the Internet to be cross-referenced using hypertext, which is embedded with references (hyperlinks) to other documents. Hypertext is placed on a computer called a web server, and users can view hypertext via client software called a web browser.

In the past, content was published on the WWW by only a few parties, such as corporations, and users simply viewed the published content. The published content was prepared in advance. Then, a system called the Common Gateway Interface (CGI) was introduced. The advent of electronic bulletin boards and chat systems enabled communication among users. The 2000s saw the spread of various technologies, and it has become easy to use blogs and social media. As a result, anyone can now publish content and disseminate information, whereas only a few parties had been able to publish content in the past. Furthermore, video sharing sites and image sharing sites have become popular. These sites are called Consumer Generated Media (CGM) because users publish content such as videos and images. There are also a wide variety of other services provided on the WWW, such as e-commerce sites and auction sites that allow users to list products.

Problems with the World Wide Web

On the other hand, WWW has various problems. I will discuss two problems in this article.

One is the concentration of access to the web server. I am sure that many people have experienced issues such as difficultly in displaying purchase pages for products that have become popular on social media, etc., or products sold on a first-come, first-served basis, videos that do not play or freeze partway, and lagging in video games. This occurs when many users simultaneously access the web server that provides the webpage. Since there is a limit to the number of requests that a web server can process at one time, the greater the number of accesses exceeding the limit, the longer the waiting time. As a result, it may be difficult to display the webpage.

Another problem is the increasing control of specific corporations. Although various people are now able to publish content, specific corporations are in charge of the platforms on which that content is published. In that case, users are subject to the policies of those corporations. Although such policies have the advantage of preventing the spread of inappropriate content, it also has the problem of preventing the free dissemination of information. In some cases, certain ideas and opinions may be suppressed. Furthermore, there are cases where a corporation's intentions are unclear, and there are cases where accounts are suspended even though there is no violation of the terms of use or the law.

Decentralized Web

In order to distribute the load of concentrated access to web servers, methods such as preparing multiple servers and distributing requests from users to each server are used. Another method is called Contents Delivery Network (CDN). CDN prepares many cache servers and temporarily stores the contents of web servers on these cache servers. Users can avoid concentrated access to web servers by retrieving content from the cache server closest to their location.

Various approaches have been taken to address the issue of the increasing dominance of specific corporations. One of the most notable developments is the emergence of decentralized social media. Also known as distributed social media, this is attracting attention as a migration destination from X (formerly Twitter). In addition to Mastodon, which appeared in 2016, other decentralized social media include Misskey and Bluesky. For these decentralized social media, servers can be prepared by both corporations and individuals, and users register on either one of the servers. It is also possible to connect servers. A group of connected servers is called a federation. This federation mechanism allows users to interact with users registered on other servers, thereby creating a framework called a Fediverse. Furthermore, protocols such as ActivityPub[1], Nostr[2], and AT Protocol[3] have been developed to build decentralized social media.

In addition to these methods, social media using blockchain has also appeared, and the trend of decentralized web is attracting attention. Although there are still technical and legal issues, I believe that the decentralized web will gradually become more widespread.

Load balancing using clients

Up until now, I had been researching load balancing using multiple web servers, but I am currently researching load balancing using clients[4][5]. The conventional web was built on communication between a web browser and a web server. In that structure, in order to share information among multiple users, it was necessary to go through the web server. However, with the advent of WebRTC, which is a protocol that supports communication among web browsers, it has become possible to share information without going through a web server. Therefore, I am testing a method to distribute the web server load to each client by building a network that uses WebRTC to interconnect web browsers. Although there are still many problems such as security issues, network optimization issues, and the unpredictability of user participation and departure, this method enables the provision of services even with a small number of web servers. The goal of this method is to reduce costs both in terms of finances and administration.

(Bibliography)
[1] "ActivityPub," [Online]. Available: https://www.w3.org/TR/activitypub/.
[2] "Nostr," [Online]. Available: https://nostr.com/.
[3] "AT Protocol," [Online]. Available: https://atproto.com/. [Accessed on Dec. 7, 2023].
[4] M. Kohana, S. Okamoto and I. Atsuko, "Optimal Data Allocation and Fairness for Online Games," International Journal of Grid and Utility Computing, 2014.
[5] M. Kohana and S. Okamoto, "A Preliminary Study of a Hybrid Data-Sharing Network for Web-Based Virtual Worlds," International Conference on Network-Based Information Systems, Springer Nature Switzerland, 2023.

Masaki Kohana/Associate Professor, Faculty of Global Informatics, Chuo University
Area of Specialization: Web Systems

Masaki Kohana completed the Doctoral Program in the Graduate School of Science and Technology, Seikei University in 2013. He holds a Ph.D. in science and engineering.
After serving as Assistant Professor in the Faculty of Science and Technology, Seikei University and Assistant Professor in the College of Engineering, Ibaraki University, he assumed his current position in 2019.
He is a co-chair of the International Workshop on Web Services and Social Media.

He currently conducts research on load balancing methods for web applications, web applications for daily life, and automatic classification of content.