Forum

Week 1: History of ...
 
Notifications
Clear all

Week 1: History of decentralization

10 Posts
6 Users
9 Likes
934 Views
Larry Sanger
(@lsanger)
Member Admin
Topic starter
 

In what ways were the networks created by original protocols of the Internet decentralized? What is (or was) "decentralized" about them? What conceptual and historical relationship is there between open source (or public domain) technology and decentralized networks?

 
Posted : 06/09/2021 3:14 pm
hampson
(@hampson)
Member Admin
 

individuals largely hosted their own privately owned content and didn’t rely on major corporations to provide basic services for web and email hosting. the open source community’s goal is to further this mission and develop tools that individuals can use to communicate and conduct business, without monitoring and censorship from corporate or government entities.

 
Posted : 06/09/2021 3:35 pm
DKoenemann
(@dkoenemann)
Active Member
 

I was surprised to learn that on early networks each person essentially operated their own server. Website views and file requests were made directly from some other person's desktop (essentially). I am not sure what I thought was happening, but I guess I never made the connection. 

In a system like this, the connection between open source and decentralization is straightforward. If the information is spread out over a large number of essentially private computers and you expect to be able to request information from any of those computers, the source code indicating how you would make such an information request of another computer and how your own personal computer would understand information request coming in from other computers, that code all has to be freely available to all. If that protocol is not openly available, the network is useless. 

Are there problems of scale here though? If I have a popular website or file, my personal computer cannot process all those requests. Server farms with high capacity (and protection) seem an inevitability once you have more than a few dozen people using the network?

 
Posted : 06/09/2021 8:18 pm
ramontesano
(@ramontesano)
New Member
 
Posted by: @lsanger

In what ways were the networks created by original protocols of the Internet decentralized? What is (or was) "decentralized" about them? What conceptual and historical relationship is there between open source (or public domain) technology and decentralized networks?

A decentralized peer to peer network is more robust, so is likely to propsper long term. Today's (2021 Big Tech) centralized internet may fade away over time.

 
Posted : 06/09/2021 8:54 pm
Christian Gribneau
(@cgribneau)
Member Admin
 

The Intenet was decentralized from the ground up, with some centralized characteristics to make things work. IPv4 addressing, as an example, uses centralized address allocation, but then many Autonomous Systems route traffic using standard protocols to determine where any given packet needs to go. The Domain Name System allocates authority centrally, but many systems participate in resolving names to addresses so we don't have to remember all those numbers.

Protocols like HTTP and SMTP (email) were decentralized by design, running atop the naming and addressing provided by IPv4 (later IPv6) and DNS. Many different servers handle those protocols, and it is still possible today to launch your own instance of each.

Back in the early 2000s, participatory peer to peer file sharing operated by users constituted a bulk of traffic traversing the internet. These systems (BitTorrent, LimeWire, KaZaA, GnuTella, eDonkey, Napster) were wildly popular but had no mechanism to protect intellectual property and the content one might find there was often riddled with various sorts of malware.

The "Web 2.0" companies, notably Spotify, NetFlix, and YouTube, became very popular by offering greater convenience, and it soon became apparent that people would pay for access to music and movies.

 

The underlying protocols are open to encourage implementation, and many of the earlier peer to peer protocols were always or have become open source. BitTorrent, as an example, is simply in the public domain.

The Web 2.0 companies, on the other hand, typically develop proprietary, closely held technology.

 
Posted : 07/09/2021 6:06 pm
Larry Sanger
(@lsanger)
Member Admin
Topic starter
 
Posted by: @hampson

individuals largely hosted their own privately owned content and didn’t rely on major corporations to provide basic services for web and email hosting. the open source community’s goal is to further this mission and develop tools that individuals can use to communicate and conduct business, without monitoring and censorship from corporate or government entities.

In the 1990s, few people ran their own servers, though it was more common for geeks to do so. What made the Internet more decentralized generally speaking is that people interacted through software that was installed on many different servers (email, Usenet, mailing lists, etc.). Yahoo was bigger than Google for many years, and prior to Google, there were many competitive search engines.

 
Posted : 07/09/2021 7:36 pm
Tim Chambers and hampson reacted
Larry Sanger
(@lsanger)
Member Admin
Topic starter
 
Posted by: @dkoenemann

I was surprised to learn that on early networks each person essentially operated their own server.

That's just not true. I think what Jean was saying (it was Jean, right?) is just that the simplicity of running things like mail servers, etc., made it much easier for SMEs each to run their own servers, which they did. That much is definitely true. Centralized email for enterprise didn't really come in until 2008-ish, I guess. ISPs, universities, corporations, etc., they all installed and ran their own email, Usenet, IRC, Telnet, SSH, and other services. The people who kept such services going were "sysadmins." Sufficiently skilled geeks ran servers right from home.

The exciting thing about NASes is that, for the first time, those services or modern versions thereof can be brought back to local management. I basically run my own replacement for web hosting (of startthis.org, anyway), Dropbox, music, images, password data (which is a relief), calendar data, etc. I could also host a lot more video (I could put my entire DVD collection on my NAS and make it all available online).

But this is for the first time. This stuff has never been doable by regular non-geeks, and NASes still need occasional support from actual geeks.

 
Posted : 07/09/2021 7:44 pm
Tim Chambers reacted
Larry Sanger
(@lsanger)
Member Admin
Topic starter
 
Posted by: @dkoenemann

Are there problems of scale here though? If I have a popular website or file, my personal computer cannot process all those requests. Server farms with high capacity (and protection) seem an inevitability once you have more than a few dozen people using the network?

Yep. Basically things got so complicated around 2000-ish that web hosting became much more popular for SMEs. Before that, again, sysadmins would keep websites up and running. If you had a lot of traffic, you literally bought more servers. Sometimes you'd pay for rack space in other parts of the world. This is still possible though. It's just a lot harder than it used to be.

 
Posted : 07/09/2021 7:53 pm
Tim Chambers reacted
Larry Sanger
(@lsanger)
Member Admin
Topic starter
 
Posted by: @cgribneau

the content one might find there was often riddled with various sorts of malware.

 

That's why I never used those systems!

 
Posted : 07/09/2021 8:30 pm
Tim Chambers
(@tbc0)
Member Admin Registered
 

Email was the first killer app on the internet, not the web. In the earliest days of the internet, online identity was strongly tied to one's email address. The most compelling reason one would have had to run one's own server would have been to receive email. It was interesting to hear Peter talk about Google's decision to offer Gmail to deter competitors. I was also amused that Jean-François chose Gmail over administering his own server. It has become much easier to administer one's own mail server in recent years. I outsource it for $60 a year, though. I remain a loyal customer of a certain ISP, which, through acquisition of another domain which you can determine without much effort, is the same service provider I started with in the 90s. I'm into personal branding, so I give out a variety of email addresses from my timchambersusa.com domain. I am privileged to have a lifetime email address through my college, and I give that address out to people I care about. Like JF, I use Gmail because it's convenient. But I don't give out my @gmail.com address to people I care about.

The web was the second killer app, and I have published through my ISP since the earliest days. I didn't care to run my own web server, either. That changed when I discovered micro.blog. I don't run it myself, though. I pay micro.blog to handle timchambersusa.com web traffic.

 
Posted : 07/09/2021 9:14 pm
Larry Sanger reacted
Share: