Decentralization ideas.

[CJJeong]

I'm Paul from Korea and very interested in decentralization. It was very interesting session today.

I think there might be some initial ideas to decentralize user's data like below :

1. Users have their own server in their home. There might be management issues because technically, users can not manage their server and the server must be weak for security from outside. So, additional management service is needed.
2. Service provider gives storage and interface to the users. Only thing the service provider can do is just to give data to other users or other services as the user(owner)'s permission. The service provider can not access to the user's data and can not use user's data for marketing or commercial. At this moment, many companies are using their user's data for commercial. But in this model, service company is strictly blocked to the user's data. Only users own their data.

And what will be the next step for making decentralization standard in W3C?

[ghost]

Hi Paul;

It sounds like what you are proposing is a combination of Zero Knowledge Privacy (models in which a service provider has no knowledge of user's content as it handles it) and federated servers (in point 2).

There are plenty of examples of what you propose: ProtonMail and SpiderOak Hive come to mind.

One major issue with these is that the service provider's interests (profit-seeking) are typically opposed to the user's interests. We still inhabit an economic bubble of "free" services where the user's information is the primary product of most consumer-facing web service companies.

Note that "federation" and "decentralization" are not the same thing; decentralization might be better described as a _con_federation of actors who agree to observe a protocol, and no more. Blockchain-based systems often function in this manner: Twister, ZeroNet, and most cryptocurrencies are examples of these.

Standardizing these systems is a hugely difficult process, and may ultiimately be infeasible or undesirable--a decentralized system or its users probably do not want a minority group capable of dictating how the system functions. Bitcoin is dealing with this issue at the moment.

"Ownership of data" as you mention is another contentious issue, spanning all the way from gene patents to user-created electronic media. Companies profit from the services they offer that employ user data as a kind of "rent-seeking" behavior; in other words, they start to be paid just for having a nice data set that is large, current, and somewhat exclusive. They guard their claims to that data quite firmly, and you can expect a fight if you want to attack those claims.

As you can see from the above, these issues are thorny and combine technical, social, economic, and political concerns. Effective decentralization as a method has to work through each of them.

[CJJeong]

Thank you for your comment talexand.

I attended re-decentralization breakout session in W3C Sapporo yesterday and I became to be very interested in this item now. Initially I just started studying decentralization at this moment and I will develop my understanding for decentralization.

Above issue I wrote is for clarifying my idea and your comment was very helpful.

[ghost]

Glad to hear it helped. Actually, your suggestions helped me organize my thoughts also. Thanks!

[threat]

I have some questions

1. Isn't a standard centralized by definition?
2. Would it be more decentralized to provide a home server to service necessity or service convenience?
3. Would it be better for one person to create a DIY version for every product ever created or would it be better if the creator of each product to also create the decentralized/DIY version?
4. Standards and specifications are used to provide integrated development, interoperability and allow for any number of implementations to be built. They are, however, often difficult to read and generally very few implementations are thus created. Wouldn't there be more competition if they were easier to implement? What other methods could be used to simplify specifications and their implementation processes? Would it be possible to convert an existing implementation into a flowchart for example, using something like jgraphx?

[AaronM04]

1. No, in fact standardization of protocols can make decentralization more likely to happen, by enabling peer-to-peer relationships.
2. I'm not sure I understand the question.
3. Not sure I understand this one either.
4. Yes, standards should be made as easy as possible to read, understand, and implement.

[threat]

1. In standardizing peer-to-peer relationships, you are also crippling potential discoveries that may prove peer-to-peer less decentralized, no? Again, by definition, isn't a standard centralized?
2. Read 1. from the initial post at the top.
3. The concept of this decentralization organization is to encourage new decentralized products and services. However, the existing products and services out there today may just need to be adjusted here and there to be more and more decentralized. Either this organization can attempt to do all of this OR this organization can encourage each person that offers a product or service already, to also offer the decentralized alternative, i.e. DIY version
4. However, the vast majority aren't and there is very little competition driving this problem to improve. How can there be a good amount of drive behind setting standards if standards organizations can't compete and be incentivized?

[KKster]

Hi everyone. My name is Konstantinos Komaitis and I am working on Internet policy issues having a law background. I am a technology geek trying to figure my way around the complex world of the Internet. I am quite interested in the whole decentralization idea and have started doing some work on it. Here is a blog post I wrote not so long ago: http://www.komaitis.org/the-conversation/decentralization-is-not-panacea-we-need-to-collaborate