Network of Networks
- 1 Introduction
- 2 User Stories
- 3 Data Experiments
- 4 Definitions
- 5 Models
- 6 Reality check
- 7 Completion
This page is for discussing and modeling the questions of how networks of networks will work. This will include refining our definitions: what is a network, what is a project? This will also include the relationship of exchange firms with a network.
This is part of the OVNI discussion.
See SENSORICA's development of the Open Alliance, which is a growing integrated and fractal open value network to which we apply the idea of networks of networks.
Network formation and organization (Sensorica)
A bunch of folks with various scientific and engineering skills decide to work together, and form a loose network to make sensors called Sensorica. They rent some lab space together. They start working on an idea for a product which they call the Mosquito. They create a website. They identify themselves as part of Sensorica when they talk to other people.
They decide that they could market a piece of the Mosquito called the Piezo. So they create a new project for that and further develop it.
Another guy comes along with another idea for a sensor not related to the Mosquito. They create a new project for that and a few of them work with the new guy on it.
They combine orders for parts and supplies when they can to get better prices.
Soon they are working on a number of projects. Some of these use the same simpler purchased parts as the others. Some of these create components that can be sold as is or used for other projects.
People bring equipment to the lab, which they sometimes share with others.
When people share the profits of something that is sold, they can make the decision of the value equation used at the project level.
Exchange firm for a network (Sensorica)
The Sensorica people know they have to have some legal entity to sell products, to be able to take advantage of benefits for legal companies, and to be able generally to operate legally in the system.
They create Tactus for this purpose. From a legal perspective, Tactus is selling the products, filing all required government forms, paying required taxes, etc. Tactus has the bank account. Tactus reports sales tax for everything purchased by individuals for Sensorica activities.
Then they decide to create a new legal organization to market other products, such as consulting for the network software. They then need to know which work was part of this and which work was part of Tactus.
It could be that other legal organizations are created that would sell products from the same projects as each other. Or possibly the same resource type. Some of the same people would work on products sold by different legal organizations.
Would/could people create an exchange firm that would represent products from different networks?
Project across networks
(The specifics of this are completely made up; but this sort of thing has been discussed.)
A few people from Sensorica and a few people from Green Wall decide to create an agricultural sensor together. Some of them work in the Sensorica lab, some at the Green Wall lab, some at home. They consider that they are working on a project together. They create a product they can potentially sell. They don't think of themselves as a long term network, but as formed more to complete a specific project, so they work together for a year and then disband.
Network of networks
Sensorica and several other organizations from the area (networks, individuals, loose associations of individuals, cooperatives) decide to come together to share space and equipment. They have some shared values. They can possibly get grant money to help with the cost of the space. This involves also sharing the space with conventional firms.
They might be able to get grant money also to buy shared equipment, which would be part of the commons of the network of like minded organizations.
The network of groups would probably need to form a legal organization to own and manage the shared equipment, and pay rent on the space.
They also want to share equipment and tools, and possibly consumables, owned by individuals or member organizations.
Network formation and organization (GMC)
Guerrilla Translators (GT) is an informal network of people that translate works they consider useful to society, they have a set of shared values. They have a website. They identify as part of the group. They don't make money at these translations. The translations they consider as becoming part of the commons. There might be incidental money coming in for these translations, but not usually.
They would like to create an agency to do translations for money, and distribute the money in a fair manner, so people can make a living at their trade. They plan to form a new organization called Guerrilla Glocalization (GG), which would do this. It would be a legal cooperative.
They also might form a legal organization for GT along the lines of a foundation.
GT and GG would be part of an umbrella organization called Guerrilla Media Collective (GMC). (Don't know if this will be a legal organization.)
Most of the same people do work as part of GT and GG both.
The actual work done for GT and GG is basically the same. The methods of distribution of money received would be different.
Network coordination (SA)
(from Ampie) There is an agriculture/food distribution network. Lots of participating farmers, most are subsistence farmers (3000+). They find that, in general, they have to transport fresh produce to urban areas.
Then there is a low cost construction network. They need to build lots of low cost houses primarily in rural areas. They find that, in general, they have to transport building materials to rural areas.
So the one network has empty trucks coming back from urban areas, the other has empty trucks coming back from rural areas.
In the typical perspective on this scenario, there is no explicit relationship between the construction network and the agriculture network. But if the building material trucks going to rural areas could be loaded with fresh produce when coming back, it may open up some really interesting opportunities.
Network splitting into multiple networks
(from Tibi) SENSORICA is about to grow beyond its initial mission (sensing and sensemaking). We are now integrating two 3D printer projects. I predict that in the future SENSORICA will split along lines of interest into two networks: sensing and 3D printing. The 3D printing design and production activity is now tolerated because capacity is low, and Serge (the 3D printing guy) is needed for other SENSORICA traditional projects and Serge needs other SENSORICA members to help him design, prototype and commercialize his open source 3D printer. As capacity grows in both domains more noise will be generated and the need to stay together will be diminished, and I predict the network to split into two.
In this reverse process I believe that the 3D printing embryo network will form its own identity and culture, as the cluster grows. At some point, it will want to express all that in a separate way, which is also good to reduce confusion about the network with the exterior (suppliers, customers, funders...). This means that the 3D printing network will want to have its own CMS.
This is a thought experiment to see how the events would flow in different scenarios. It assumes agents can be people, exchange firms, network communities. (Don't know yet how projects fit in.) The purpose is to see if and how the event flows might work if we don't have the context of Sensorica - that is, if there are more than one network, more than one exchange firm, projects across networks, etc., in one system.
|From Agent||Resource||Event Type||To Agent||Context|
|Francois||20 hours||work||Tactus||Non-production work|
|Alibaba||Electronic Parts X ($300)||ships||Tactus||Exchange 1|
|Alibaba||taxes ($23)||expense||Tactus||Exchange 1|
|Tactus?||Electronic Parts X (some of them)||consumed||Tactus||Process 1|
|Jonathan||15 hours labor||work||Tactus||Process 1|
|Tactus||Electronic Component 1||produces||Tactus||Process 1|
|Tactus||Electronic Component 1||consumes||Tactus||Process 2|
|Ivan||15 hours labor||work||Tactus||Process 2|
|Tactus||Piezo||sells||Customer 1||Sale 1|
|Customer 1||$3000||pays||Tactus||Sale 1|
|Sensorica?||$900||pays||Supplier 2||Exchange 2|
|Supplier 2||Computer||ships||Sensorica?||Exchange 2|
|Crowdfunder Joe||$100||contributes||Robotics Lab Co-op||Exchange 3|
|Robotics Lab Co-op||PID*||?**||Crowdfunder Joe||Exchange 3|
|Crowdfunder Mary||$100||contributes||Robotics Lab Co-op||Exchange 3?|
|Robotics Lab Co-op||PID*||?**||Crowdfunder Mary||Exchange 3?|
|Tibi||2 hours labor||work||Robotics Lab Co-op||Non-production work|
|Robotics Lab Co-op||PID*||?**||Tibi||Non-production work|
|Robotics Lab Co-op||$300||pays||Landlord A||Exchange ?|
|Tactus||Electronics Parts X (more of them)||consumed||Robotics Lab Co-op||Process 3|
|Antonio||5 hours labor||work||Robotics Lab Co-op||Process 3|
|Robotics Lab Co-op||PID*||?**||Antonio||Process 3|
* PID = Participation in Income Distributions.
** Event ? means PID might be implicit, in other words, no balancing event will happen at the time of the contribution, but the participation will be assumed. Or they might be explicit, in other words, a balancing event will happen, which will need a name, as will the PID, which will need to be some kind of Resource Type that is given in return for a contribution.
- Agent - A person, legal organization, or informal collaboration (like a network), which participates in economic events.
- Network -
- Project - See project page
Some brainstorming on what we want and how it would work, thoughts are evolving. Where we are now technically is that one instance (of the app and database) assumes one network (Sensorica or GMC). The data assumes the network in a lot of places, just like in enterprise software, the enterprise itself is always assumed. We had originally thought that the way this would evolve was that new networks would implement new instances of the software, and would communicate through a protocol (of necessity). This also will likely be needed in the future, but it has become apparent that in and around Sensorica, the evolution is much more fluid and complex than this. (See user stories above.)
- Support network of networks (Open-alliance)
- Support multiple exchange firms in Sensorica
- Support Robotics Lab (in whatever form it takes)
- Support Guerrilla Media Collective using the software developed for and with Sensorica (GMC will live in a different instance of the software than Sensorica)
- Make one instance of the software / database support any combination of related networks, people, legal entities (as exchange firms, as the network form of organization, whatever). Two implications: the Sensorica installation will be broadened beyond Sensorica; we will support multi-party accounting either within one network or for multiple networks. The general implication is that as things morph in and around Sensorica, people and groups can fold themselves into this installation of the software without the obstacle of setting up a new installation.
- Figure out and clean up the question of how projects fit into the scheme (not every network needs projects; in Sensorica, projects have some of the same behavior as networks and some of the same behavior as internal work orders). It is not clear if projects should remain a distinct concept, or become agents and in a few cases become internal orders.
I don't know yet if we can do the first without resolving the second, or not. Need more discussion and conceptualization around projects.
Model and UI
The REA model supports multi-party accounting and in fact any set of relationships between agents through economic events. The OVN-OS database also supports all of this at the core. We finessed this in the actual data because we didn't need to instantiate Sensorica or figure out how Tactus fits into economic events and reporting because they were the only network and exchange firm there. As well, there was some discomfort with thinking of "agency" in relation to Sensorica, still something that needs to be discussed.
The core model supports all of this nicely, but we will need more definition at the type layer around agent types and agent roles in relation to each other. (See agent model in this section.)
The UI does not input agents for events in cases where Sensorica or Tactus is assumed.
What this could look like
So what are we talking about concretely? Brainstorming:
First, we need to input or derive the from-agents and to-agents in all events. Then accounting or government reports can find the correct events for the entity doing the reporting (like Tactus). We can set up structures so in some cases the event agents can default, which can then be overridden or not (a decision would be needed). For example, projects or networks (or orders??) can assign an exchange firm to themselves. Then the events could be recorded or reported out against that exchange firm for work done in the project or network (or order??) The question of recording or just reporting involves timing of the assignments. Do we want to be able to decide at the last minute? Is this even possible? Do we want to record all the agents on the spot in case relationships change later and we don't want them to be retroactive? (The latter is my current opinion, but need input.)
Note if projects become agents, Sensorica can be thought of as the highest level project in the current scheme. Although really you end up with a hierarchy of agents, which in this case are loose groupings of people working together on something, perhaps a product line. We already know that current projects want to be able to define their own value equation, and perhaps elements of other governance. This is agent-like behavior. These "project agents" could choose to not define value equations etc. for themselves, in which case, they would inherit these from the parent, on up the chain. Membership could work that way also. If projects become agents, projects don't have to explicitly change into networks if in fact they do move outside of the parent network.
How to treat events that create or bring in resources that become part of the commons? And are there different levels of commons or custodians of shared resources or ?? Like, someone can donate a microscope for use of anyone in the Piezo project, or whatever....? How do we want to account for these in the events? (We already have set up some associations around the resources themselves (owner, custodian, etc.), which are related to this question, but currently explicitly defined in the data model, not derived from the event that created or contributed or purchased the resource.)
In Sensorica, do you want to think of the project as autonomous enough that it can decide when to allow resources it created to be used in other projects?
Users will need different permission levels for different networks.
Users could potentially pick their network context among the ones they are member of, and this would default down as they log etc.
Agent (organizations) would be an additional filter on: All Work, Network diagrams, ?
Internal work orders may need to be more prominent, as they may take the place of existing low level projects. (Some Sensorica projects should already be work orders anyhow, I think.)
An idea was discussed between Lynn, Francois and Tibi, to give networks the ability to nest-incubate within a network and be spun off as independent networks after. The Drone and 3D printer projects have the ability to become networks. At this moment they are mostly using SENSORICA's resources, so they are developing as an embryo within the network. This will most probably repeat itself, since new networks need help and a platform to start their life. This reality tells us that OVNs need mechanisms to incubate other networks that will be spun off later.
Growth of network of networks
We observed in SENSORICA 2 growth modes for Network of Networks.
- by affiliation - independent networks that already exist affiliate with other networks. See work on the Open Alliance.
- by incubation and spin off - individuals with project ideas join a network, grow their projects which grow into networks, use local resources to develop and are spun off once they reach critical mass. This is the case with Drone and 3D printing projects in SENSORICA.
We are thinking about giving OVNs the internal mechanisms to nest and incubate embryo networks to be spun off later.
As of 5/2014, we completed implementation of these ideas in the base valnet software. This means that the models and design discussion above are now out of date, and for correct models, people should refer to the software itself. (This is always the case. It is not worth keeping analysis and design docs up to date.) However, the user stories above remain valid and useful.
Some related requirements remain - accounting reports and exports; permissions for visibility to part of the data based on agents one is associated with (if this is a requirement, unknown right now).