Forum

What do we need mos...
 
Notifications
Clear all

What do we need most, to create a decentralized network?  

Page 3 / 3

dannotestein
Posts: 1
(@dannotestein)
New Member
Joined: 1 week ago

We have been working for several years on a decentralized social media platform called Hive. It operates on a peer-to-peer network and uses a blockchain to store the text data and the URL hash links to other forms of data (images/video). Users create accounts that are tied to cryptographic keys that allow them to control their account information.

The data can be accessed via a number of different web interfaces, here's a few:

I have to go now, but I'll leave more info later about how it works, and some of the plans for the future.

Reply
4 Replies
marcabisaleh
(@marcabisaleh)
Joined: 1 week ago

New Member
Posts: 3

@dannotestein I have been a hive user for years now. No solution is perfect but this by FAR the best one I have found. And I searched a lot.

Reply
seriouscoderone
(@seriouscoderone)
Joined: 1 week ago

Active Member
Posts: 6

@marcabisaleh What are the issues you have seen with Hive?

Reply
marcabisaleh
(@marcabisaleh)
Joined: 1 week ago

New Member
Posts: 3

@seriouscoderone I have written a long reply just below, with a lot more information on hive. I'll write a recap of the biggest issues I have seen: 

-inability to store images and videos (it is text only). This is mitigated by posting URLs to IPFS databases for example. 

-No privacy. It is extremely complicated to implement. Accounts on hive are pseudonymous, so if you are capable of remaining anonymous on the internet, it is possible to use Hive anonymously. But this is too hard for most users. 

-the governance inherently depends on Hive holders. Larry Sanger raised that point on twitter, talking about whales (large coin holders) dominating the protocol. This is mitigated through forking. Before becoming Hive, Hive was called Steem. It was attacked when a large stakeholder used his majority stake to censor posts. The community simply forked away from Steem, which has become a dead centralized chain. The stake is now extremely well distributed on hive, with the largest whale only having a 3% influence on the network. This of course can change in time, but every attack on the network is expensive, and we can always fork. I highly recommend this article if interested in how to survive attacks: https://decrypt.co/38050/steem-steemit-tron-justin-sun-cryptocurrency-war

-Early development. Tokenization has to be enabled for communities, and the whole UX can be better. This will happen in time.

Hive is complicated and can be better understood when using it. https://www.hive.io

Reply
marcabisaleh
(@marcabisaleh)
Joined: 1 week ago

New Member
Posts: 3

Here is a tweet thread I posted in attempt to correct some misconceptions about Hive. @dannotestein is the expert, and I am just a user of Hive, but while we wait for further posts from him, this might serve as a good start: 

1/ I understand your point of view, but you are missing some things! I will try to answer your points, and provide an objective overview. Hive is definitely not the perfect solution, but it should be given more credit as it is not victim to some of the points you mentioned.

2/I'll start with concessions. Hive can't store images or videos in a decentralized way, but simply text because it's cheaper. For videos, and images, the solution would be to use something like IPFS or bitTorent and then add links in hive posts.

3/Hive is not private, but pseudonymous. If a user is capable of remaining anonymous over the internet, he is capable of using hive without sharing his identity. Obviously, this is too hard for most people.

4/Now to the positives. Hive is definitely not a website/platform. The best way to think of it is as a layer 1 protocol that enables decentralized text storage. On top of that are second layers (or sidechains). Currently, the most successful second layer is Hivemind.

5/Hivemind is a sidechain that tries to decentralize social media. It is currently accessible from multiple frontends: such as ecency.com, hive.blog, and peakd.com. Ecency and hive.blog are open-source.

6/Both hive's implementation (hived) and the sidechain (hivemind) run on multiple nodes, in a typical blockchain fashion. The number of nodes is still low, but it is early days 🙂
Hivemind also allows the creation of communities.

7/Communities have their own governance. The vision is to have tokenized communities. In each community the leader/founder/owner can curate and moderate the content. If he censors too much, users can start their own communities and move. Think of subreddits, decentralized.

8/Communities already exist, but the tokenization is still in development. Each community will have its own whales dominating, but this problem is solved by the ability to create competing communities. In the end the users are king, and a whale must not drive users away.

9/As for the governance of the protocol itself (hive layer 1), it is also stake based and can be dominated by whales. In fact, Hive did have this problem when it was called Steem. I highly recommend this read: https://decrypt.co/38050/steem-steemit-tron-justin-sun-cryptocurrency-war

10/The article shows the power of forking. There is always a possibility of the network being overtaken by a huge stake, but the result is simply a dead centralized chain dominated by the attacker, while the whole world shifts to the new decentralized chain. Steem --> hive --> ?

11/Attacking the protocol requires a lot of money, as the attacker needs to become a whale and/or convince other whales to censor with him. As soon as the attack succeeds, the community forks away. The chain can be attacked infinitely, but each time attackers wastes money.

12/Once we have that security, any kind of layer 2 sidechain can be built on top. Right now hivemind is the only option, but with communities it allows every community to have the censorship, governance, and stake distribution it desires. @FinanceLeo is an example of a community.

13/@FinanceLeo have their own token and different whales than hive. Since the technology for tokenization is not developed yet for communities, their token is currently semi-centralized (transactions posted on hive). They are developing their own layer 2, in order to decentralize completely.

14/ I am simply a hive user with limited dev skills. There are lots of much better people to explain than me, here are some of them: @blocktradesus @FAarrestad @ausbitbank @therealwolf42 and many, many more whose twitter handle I don't know. You can find everyone on hive.

15/I will also write a post on your new blog with the same info I used in that thread. I hope this gets you more interested and that you can contribute to improving hive. I truly believe it is one of our best shots at decentralizing social media and freedom of speech.

Reply
tkdcat
Posts: 1
(@tkdcat)
New Member
Joined: 1 week ago

Decentralizing the application and infrastructure is doable but can be complicated. The real trick will be decentralizing the DNS. ICANN is now a private company. You need to register a domain name with a registar to be found on the internet. At some point, ICANN and the registar will simply deny allocating a public IP address for the application. This will make all these other points moot. If we are going to decentralize we need to find a way around the private company that controls the Internet's IP address allocations.

Reply
ProteusMirabilis
Posts: 2
(@proteusmirabilis)
New Member
Joined: 1 week ago

For several years, my colleagues and I have been researching and developing a peer-to-peer system designed to replace HTTP as the primary protocol for the Internet.

More recently, several (entirely separate) projects sprung have emerged from the ether with similar motivations, namely:

  • Secure Scuttlebutt
  • DAT
  • IPFS

In a free and open Internet, any of these protocols can be used to facilitate secure, decentralized storage and network communication. However, in the context of a well-funded and hostile adversary hell-bent on censoring communications, each of these technologies have serious weaknesses when it comes to:

  • User Anonymity
  • Peer Discovery
  • Network Filtering
  • Data Dissemination

Without going into too much detail, our research led us to conclude that the design of Secure Scuttlebutt (SSB) is closest to meeting our objectives, with modifications to address the weaknesses above by adding:

Discovery

  • Proximity-based (Bluetooth, WIFI)
  • Broadcast-based (UDP, Multicast)
  • DNS-based
  • IP-based

Encryption

  • User identities
  • Feed identities
  • Connection encryption
  • Payload encryption 

Dissemination

  • Opportunistic gossip-about-gossip

Persistence

  • Mutual reciprocity
  • Proof-of-Storage

Deception

  • HTTP/TLS overlay protocol
  • Firewall hole-punching

Our work to date has been nascent, and we have not yet opened the project to others outside of our core team. Now, given recent events, it seems we were foolish not to create an open development community at the outset. The hour is late.

With that in mind, I’d like to strongly suggest that SSB is a shortcut to where we would all like to be  technologically, and encourage those who are technically minded among us to join the development of SSB directly.

We intend to merge our progress with and join SSB in the near future. I expect our work will be more than welcomed, but in any event the license of SSB is broadly permissive (MIT), and therefore we needn't worry about our ability to pursue our ends separately if necesssary in the long term. 

 

Yours,

Proteus Mirabilis 

Reply
narby
Posts: 2
(@narby)
New Member
Joined: 1 week ago

After reading some posts, it is apparent a few groups of developers have done some stuff.

That’s great. But.

In the end, the weak link with any set of code intended for wide use is the marketing. Getting the system/standard into wide distribution. I assume there may never be a large monetization capacity here, so getting the system into wide distribution will be difficult to organize. 

The subject of this thread, “What do we need most to create a decentralized network”, I believe the top priority is a plan to get this spread across the world. Indeed, the value of a “social network” is in getting a sufficient number of people involved, and in enabling the network to self-organize into smaller sets of individuals that have common interests, and maybe more important, no preexisting hatred to divide them.

In observing how the large socials operate, it is apparent they have some very highly developed software designed to hook up individuals, and tamp down fist fights. This sort of thing might be impossible to develop in a decentralized network, but may be possible by designing the network architecture as a standard and allowing any number of secondary software organizations to write code to use the system.

That paradigm will allow the “smart” element of building a strong and distributed social to be at the end points. The user interface level of the system. And those user interface elements should be easily replaceable, just as I can use any number of different web browsers to access the same sets of URLs.

Reply
1 Reply
Killface
(@killface)
Joined: 2 weeks ago

New Member
Posts: 2

@narby This is a good point that needs more attention. This may sound obvious - it's not and is often overlooked - but the first step to marketing the network depends on ease-of-use. The end-service must to be so easy to adopt that my grandmother could do it in 3 steps. If it's too complicated, the service becomes another niche hobbyist domain. Simplify. Brevity is better. Remove friction or loose.

Reply
darenwelsh
Posts: 2
(@darenwelsh)
New Member
Joined: 2 weeks ago

I haven't read the article or discussion but figured it might be relevant:

https://news.ycombinator.com/item?id=25731419

Reply
Page 3 / 3
Share: