Is Internet's Centralized architecture a problem?


Centralized architecture in Internet refers to the scenario where data is stored in a one or multiple location but they belong to single entity. For example, Everyone knows Facebook or Google runs a number of data-centers across the earth and thus our data is stored in one of their data center.

Now that this fact is known, there is no surprise that anyone who has authority or power can approach Google or facebook to get users data. Just like how NSA or Advertising agency gets it. This is not just with google or facebook, this is how the web today functions.

If you think this centralized architecture is flawed, why do you think so? What can we do about it?

We should move away from Telegram

I think centralizing the data is not a problem, because only those servers can handle that much data. The problem comes when their source codes are not open. Even if data is centralized but the source code is open and the source code is verified that it does not steal our data, then i think it wont be an issue. Please correct me if i am wrong.


In my view, I can’t come to a conclusion that the system is flawed. We just can’t come to a conclusion just because our data is with someone else. If that is the case, consider our money with the Bank and all our earnings are represented in 1’s and 0’s. Died Hard 4 clearly depicts what could be done to the financial system when everything is connected. Whatever system we design there will be issues because they will suffice one and not the other.

Yes, there are issues with Centralized architecture and so does Decentralized system. Centralized systems can have multiple backup while in decentralized it can be done but more expensive. Participants in a decentralized system should actively engage so that others in the system are well benefited. For ex: if there lesser number of seeds for a torrent then the overall efficeincy is much lower. To quote advantages of decentralized system - there is no single point of failure, better computing power and so on.

This is really a healthy topic dude…@pras


@arvinth - being open and not open is onething. Even if Facebook was open, when its under some govt. control its being forced to provide data which is the main problem that @prashere has highlighted.


@vms20591 There is a basic form of invisible trust between a User and the Service provider, that the user subconsciously trusts the service provider. I am not at all blaming the service provider here.

My point is that, the centralized architecture by itself is attracting Governments / Private agents towards the huge pile of data stored. These agencies now know there are this number of user data and they also know from where to get it, so they just go and grab the data required.

Not even required / targeted data, they wanted to get as much data as possible, so that they need not depend on those providers for sometime. I state the problem here is they are approaching the service providers because agencies know where the data is because the data is centralized.

They could easily manipulate everything which affects not just the privacy of an user but also affects the economic growth, political nature of a region, public sectors, the progress of a particular society and many more.

Exactly! I believe you are aware of what Bitcoin is trying to do.

Again, I completely agree with you and that is what we want and the system has already worked out. The success stories of Napster music streaming service, Torrent file sharing service. And by the way, when the system like torrent just works and people do not care about the ideology behind the peer-to-peer sharing system but still they prefer using those services because of the fact it just works for them.

We should also note that the above two services reached more number of people by being centralized for just node lookups. They can also be decentralized and for that to be accessible by more number of people like in centralized system, there are researches undergoing like the project meshnet.

I strongly believe peer to peer and decentralized systems can make through if they can handle about any scenario we throw at them.

@aravinthk I agree we cannot trust a closed source or a proprietary system that says we are not doing any mischief with your data. As the famous quote goes like this

Talk is Cheap. Show me the Code.

And moreover, the primary point here is Centralized vs Decentralized than Open vs Closed. But yes, the nature of the source matters along with the architecture too.


Centralized architecture is always a problem…People who control the internet and user data can even control the country…Privacy is not the only issue with the centralized system, it is much worse than what we think… If internet is controlled by centralized parties it will become the most powerful weapon…@prashere This topic is more sensitive man


Excerpt from the above article

“The simplest outcome is we’re going to end up breaking the Internet,” said Google’s Schmidt. Foreign governments, he said, are “eventually going to say, we want our own Internet in our country because we want it to work our way, and we don’t want the NSA and these other people in it.”

I don’t think they are breaking up the internet and nor it will happen.

What is happening is an natural Evolution of Internet. The Internet will take a new form and that form will occur only when each one of us help the regular Internet user to understand the underlying problem, making things easier for them to use and spread the ideas of communicating securely.


@vms20591 @aravinthk @rajanand and all check this out. Interesting blog post on Decentralized architecture.


Thats a nice blog :slight_smile:


@vms20591 @aravinthk @rajanand
Check this out.

Interesting project.


Thats a nice one !!! :slight_smile: :slight_smile:


I am a web developer, thanks for this article, gave me a sneak peak of what would becom the future.


Lets say I get the source code thats under GPL, then modify it on my server so that it analyzes the data and forms patterns about people and you can see the sourcecode thats run on my server, then how can you say my server is clean?

Centralization is huge problem, the data can be duplicated on another hard disk and sold to the highest bidder. Decentralization is the only way to go.


Decentralization + Peer-2-Peer.


Good read @prashere.


P is for Peer-to-Peer Networking

P is a small framework used to create browser-to-browser networks (as opposed to just a connection).

With P, you can:

  • Connect to other browsers using a simple WebSocket server.
  • Connect to other browsers using your established connections to other browsers.
  • This is what makes P unique, it allows for transitive connections across peers, allowing easy creation of mesh networks.

After a connection is established the middleman is no longer nescessary – no proxies are involved.


consider a current state of skype and other simliar communication solution , they are heavily centralised, hence there is a huge latency on both end. On other hand, communication based webrtc standards works best where latency is low since communication depends directly between the computers and server is only responsible for discovery.


here list of resources on how internet can be decentralised


Skype uses peer to peer communication, where as Google hangout uses client-server architecture.


Google early versions used client-server but new version don’t use it anymore.
, even though skypt says p2p we can’t be sure its a proprietary protocol.