Incentive Network Protocols

From P2P Foundation
Jump to navigation Jump to search

= "the application of economic principles to web development".



Tiberius Brastaviceanu:

"I recently discovered a body of knowledge "incentive network protocols". In short, this is the application of economic principles to web development.

More and more software is not just software anymore, dealing just with information processing. Web infrastructures have now baked into them economic models. Code users like you and I are not just passive information consumers, we are at the same time interested and motivated and active, we are contributing in some way by rerouting information, storing and providing data, etc. This is mostly true in p2p infrastructures. The best example is Bitcoin. All previous attempts to create a p2p digital currency have failed, until Bitcoin incorporated economic incentives and costs in order to model users' behaviour, to secure the network. It turns out that p2p filesharing, where p2p actually started, is now also benefiting from the same strategy, merging economics with informatics. Here's an article that illustrates this.

But why should we be interested in how computer network protocols work? Sensorica is not building such things. Well, in my opinion there's just one step to understand how sensoricans operate from how people sharing files online operate. Tweaking the p2p file sharing protocol to take into consideration the fact that some peers act altruistically and others engage in free riding, in order to optimize the network's performance for searching and retrieving files, sounds a lot like tweaking Benefit Redistribution Algorithms based on data from the NRP-CAS (network resource planning and contribution accounting) to allocate benefits in proportion to contributions, in order to improve the creativity and the productivity of the network.

Unstructured p2p filesharing networks that don't use incentives or reputation develop inefficiencies because of a high level of free-riding. The altruistic peers become overloaded with requests and pay a huge price, known as the tragedy of the commons. Lots of people with good intentions, people that I like, believe that in organisations like Sensorica everyone should have equal access to everything, because somehow "love" is going to solve everything. Then why do we have so many people that burn out in open networks and communities?! I would love to live in a world where everyone has access to everything, and everything works just fine. But that world doesn't exist and I don't think that humans will build it. There are free riders and opportunists in this world and they are not just a product of our current society. There are psychopaths among us, nothing to do with our upbringing and culture. These people and others of similar kinds, if we attract a critical mass of them and let them have it their way, they bring our organisation down. Sensorica had a fair share of them in the past, just search the word "crisis" in this mailing list.

So what can we do with this? How do we translate it into governance, infrastructure and methodologies?

For one thing, we need to congratulate those who have contributed to the OVN model, because we're definitely on the right track. I am still trying to inform other open networks who are still trapped by the belief that if we try to be good and friendly we'll influence others around us to be good and friendly and we'll be able to scale a good and friendly network. It turns out that that thing doesn't scale, especialy as an open network. The best we can do, is to create a well guarded small community with a powerful filter at the gate to let in only those who can prove their natural goodness and friendliness. In other words, these communities collapse once they open their gates and get the visit of only a few psychopaths and opportunists.

Second, we need to continue to refine our models of benefit redistribution. What are the benefits provided by an organisation? Not just tangible or material, but also intangible such as visibility, recognition, sense of belonging, etc. One important thing is the balance between the degree of openness (access to participation and hence to benefits), which is vital to a network to replenish its ressources, the level of transparency (access to information, which can also be seen as a form of benefit - learning, knowledge, connections, opportunities, etc.), which is a necessary condition for openness to be effective (you decide to participate based on your knowledge about the thing that you want to join) AND, on the other side, limiting access to benefits only to those who merit. In short, how can you maintain an inflow of participation, based on transparency and openness, knowing that the same openness and transparency gives unregulated access to some benefits. These benefits that are provided without restrictions cannot be used as leverage to model the behaviour of agents. So we have a paradox.

Here's an example. Sensorica provides unrestricted access to documentation, to everything we do in projects. Open source has proven that it is precisely this transparency that drives in participation, which sustains the project. But that same transparency allows people passing by to learn something new (benefit) without giving something back to the community. The argument in favor of transparency goes like this : it's OK if a billion individuals benefit without contributing as long as a critical mass of individuals do contribute, enough to advance the project. The problem is that over 90% of open source projects function way below that critical mass and are as sustainable as the tolerance to burnout of the very few in charge. That argument holds only in special cases, such as Linux, Mozilla, and a few others, which generate huge flows around them, capturing a very small portion of it, but enough to sustain them. What about all the others? Should the other 90% of open projects adopt a mixed strategy, ask you to earn your right to view everything? Is that going to increase or decrease participation? I suppose it will increase participation only if the organisation has something very special to offer that cannot be found elsewhere.

Should we put a "proof of work" wall in front of our documents?

Should that wall be higher or shorter based on the type of information? I suppose that depends on our development strategy, on the role we want to play in our ecosystem." (email, December 2019)