An alarming trend of internet in the past few decades is the fast concentration of traffics on fewer apps or sites owned by tech giants, who used their dominant control of data to influence, even manipulate user behavior for monopolistic profits, at the expense of average users. However, we hold the following truth to be self-evident, that a free and open internet is of the best interest to its users.
Leither, a decentralized Cloud OS, is designed to take the challenge by decoupling and reconstructing most of the core services of regular internet in a decentralized fashion, and give the control of data back to user, its real owner. Leither OS has implemented all the subsystems necessary to reconstruct regular internet services, such as authentication and authorization, file system, database, application framework, domain name service and routing.
To reconstruct a regular internet application, for example Youtube or Google, simply creates a Leither Organization and a Smart Contract to stipulate how to distribute tokens among participants. Token is a proof of contribution that guarantees proportional dividends of the organization's income, or can exchange for services such as traffics, routing, domain name resolution, search, backup, bandwidth, CPU power, contract notary, or mining pledge.
Members of an organization include founder, investor, developer, content or service provider, device operator, operation and marketing, etc. The distribution of income is guaranteed by consensus. Device owner does the work of miner in BTC. Devices keep the books as mining machines in BTC, at the same time provide more valuable business services of the organization. Compared with POW-based block-chain, Leither has advantages in all aspects of cost, efficiency, performance and privacy, and last but not least, law abiding.
This paper has two parts:
Part I: Introduction of MiMei and its application system.
Part II: Organization and Consensus.
How to decouple and reconstruct regular internet platform services.
Definition of key concepts:
-
Data
Information stored on any media. Data structure defines how data is stored and organized. It is a set of data items that collaborate with each other following certain rules. -
Application
Data must be consumed or processed by program. There are two types of programs in general, system software and application program. The latter is used to process user data, aka APP. -
MiMei
The key concept of Leither, inspired by Gene and MEME. MEME is the gene of civilization, an element of a culture or system of behavior passed from one individual to another by imitation or other non-genetic means.MiMei is the carrier of information on internet. It defines the structure of data storage and operations by associated applications, such as create, open, save, read, search, render and user-interact.
Everything in Leither-based ecosystem is defined in term of MiMei. The referential relationships among MiMei makes it possible to develop comprehensive internet applications and services.
-
Leither
is a container of MiMei, an ultra-light decentralized cloud OS.Leither is designed to support frequently used internet applications with extremely limited resources, for example to run office applications for a small business on a Raspberry Pi Zero.
Leither OS has implemented all the subsystems necessary to reconstruct regular internet services, including authentication and authorization, file system, database, application framework, independent domain name service and routing. Leither API supports about 40 kinds of development languages.
PS: Leither written in Kanji character is 弥媒. 弥 means extremely small and ubiquitous, usually something imperceptible. 媒 means media. 弥媒 put together is almost as good as Ether, a substance once believed to be the building block of universe.
-
Node
A smart device running Leither OS is a Node. As small as 6MB in size and runs on all regular operation systems, Leither runs almost any smart devices, like PC, NAS, router and Raspberry Pi.
-
One independent node can serve as an ultralight server;
-
Two nodes can backup each other in services like data backup, error tolerance, load balance, domain name resolution and routing;
-
Multiple nodes can network to support decentralized internet applications.
MiMei flows among nodes to facilitate the decentralization of centralized internet.
Internet has changed our daily lives. Big Data is running the internet, therefore our lives. Online retail went up to US$35 trillion in 2019 in China alone.
The essence of Big Data is the accumulation of data and control of rules, for the purpose of manipulating user behavior and generating monopolistic profit.
Big Techs own core data. For example, FB owns social media and Amazon online retail. Data protocols are manipulated to serve as business barriers. Google’s spider cannot access FB pages, vice versa.
Monopoly is protected by corporation regulations, laws and ecosystem rules. Big Techs also take the lion’s share of profit, so their fortune concentrates accelerating.
Individual user is helpless against the invasion of Big Data into its privacy. All user behaviors are data tagged to feed data models, which is then used to predict and influence user behavior.
Small business struggles under the piercing peek of Big Data, to which the ledger of small business is an open book. Small game developers can barely share 10% of income on Tencent Game. Most online retailers are hard pressed to break even.
The internet originated from ARPA net of US Army as non-profit project. It was borne an open system without a center. Telecom firms have invested billions of dollars to build infrastructure, without controlling the data flowing through it. However, Tech Giants, such as Alibaba, Tencent, FB, Google, etc., has taken control of internet with a few APPs.
The internet of Status quo has become a centralized playground of big players. By controlling user data and influencing user behaviors, monopolistic profit is being made at the cost of individuals, for whom a weapon to fight back is in dire need.
Before presenting a plan to build decentralized internet, it is necessary to briefly review the different requests that computer and internet must address at each stage of their development, and the corresponding solutions presented. If the new solution cannot solve old problems satisfactorily, it cannot beat the incumbent.
In the beginning of computer age, applications were developed to solve scientific problems. Those application generated the earliest request of data, and also large amount of data in return.
In Unix, File was designed as carrier of data. File System was designed to manage files and storage medias.
Everything is file. Any information related device is of File type.
With the evolution of applications, it became necessary to describe the relationship between file and application. DOS introduced the concept of File Extension. In Windows, file extension is bound with a corresponding application. MIME defines a strict rule for file type as E-mail attachment. All existing web browsers still follow the MIME protocol.
Database was designed to manage complicated data structure. A database is a closed space of datasets and their relationships.
Shortcomings: Data stored in closed space with little liquidity, value not maximized.
WWW regulates connections of data cross machines over a network. The inter-connection of data cross network breaks the physical limits of data storage on mainframe. With the birth of internet, information explosion ensued and gateway sites flourished.
Shortcomings: connections based on linkage is vulnerable.
Information explosion creates the request for search engine. By adding tags to data, Google provides content search service by keywords.
Shortcomings: Data is more and more isolated and monopolized. Data protocol began to be influenced and controlled.
The early social medias, ICQ, MSN, QQ, solved problem of instant messaging among people. Twitter provides a platform for broadcasting short messages quickly.
From MySpace to Facebook, a more systematic social media model is created. The consists of 4 parts:
| # | Title | Detail |
|---|---|---|
| 1 | My settings | Personal information, settings and preferences |
| 2 | Friends | User contacts |
| 3 | App & Services | Application and services for the user |
| 4 | Message | Interactions with friends, applications and services |
This model is remarkably successful. All the afterward internet social Apps adopted a similar framework, for example Wechat, Facebook, Alipay and some OA system for enterprise users.
Shortcomings: Internet becomes more monopolized, more isolated, and data protocols manipulated.
Design principles for breaking data monopoly and creating decentralized internet:
- User oriented in every aspect
A new birth of freedom, and that internet of the people, by the people, for the people. - Free and open system
Open data protocols.
Open operation system protocols.
Only in an open system, user has the freedom of choice and power of control. - Free Data Migration
With user’s authorization and reasonable security discretion, user data shall be allowed to move from one platform to another and transform from one format to another. Only then, a third party can provide valuable applications and services.
The following requests shall be supported in a decentralized system:
- File, file system, database
- Association of data type and application
- Association of contents over network
- Search of content over network
- Data migration over network
- Decentralized social model
- Light-weight container system
- User friendly development framework
- User experience compatible with existing internet
- Law-abiding
First and foremost, an understanding of how data is monopolized.
In nascent internet, data access follows open protocol across the whole network. Major platforms like Google and Microsoft, all provided open API for accessing their data.
The earliest abuse of open data policy was committed by Facebook, which used other platforms’ data to construct its own social media, then shut the others out. User data get stuck in Facebook’s platform and the first monopoly of data was borne.
Afterward Apple and Amazon created their own kingdoms of user data, respectively. When Tencent published QQ, it used a proprietary communication protocol. When Baidu blocked commercials of Alibaba, the latter responded in kind.
Thus, internet became segregated. User data is fortified by impenetrable barriers created by every monopolist who can get its hands on it.
There are two types of data access.
Data->App
User can freely access data, then choose whatever application to process the data. For example, in Windows user can select different applications to open certain type of file, or use different browsers to open a web page.
In this case, User has the freedom of choice and is not restricted by certain App.
App->Data
User's right to access data is restricted to certain Apps, usually only one. For example, when user login an e-commerce site for shopping, or get on social media to chat with friends.
In this case, the algorithm of App decides the path for user to access data, so user behavior is controlled by the App's algorithm or rules.
The two approaches above correspond to Object Oriented Programming and Procedure Oriented Programming, respectively.
Open Source Platform
Microsoft dominated the PC ecosystem with DOS and Windows, and strangled competition in both office software and browser. Google uses search engine to control data and influences user behavior. FB does the same with social media.
On the other hand, open source community turned operation system into a public platform with Linux. It is the first attempt to open source a platform. Similar cases are Android for smartphone and RISC-V for IC instruction set.
There shall be similar open source platform for collecting user data, especially for Social Media and Search Engine.
Free Data Migration
In early operation system, data file can be copied and transferred among different mainframes of different systems. Free migration of data shall be a consensus and primary design norm of operation system.
Three measures can be adopted to achieve the above goals:
- Promote data-freedom as a mainstream ideology
- Mandate that any platform that collects user data must allow the owner to migrate its personal data. (European Commission proposed a similar legislation recently)
- Develop open source programs, such as browser plugins, to facilitate user to migrate data to personal storage devices. Once the ownership of user data taken back by its creator, the data can be revived and reconnected with each other.
Rule subordinates to data
If data can be migrated freely, user will have the choice of rules to apply to it. In another word, try to be Object Oriented, avoid Procedure Oriented development whenever possible. Data structure is relatively simple to be reverse engineered. It is easier to reconstruct program structure with the knowledge of data structure.
Granulation of User Data
There is no way to decouple convoluted and complicated dataset. It is best to granulate data in design, following the pattern of gene, keep the information relatively comprehensive and independent. A piece of granulated information is easy to be processed, transferred, and its connection with other objects clear to be understood.
inter-Connections of Data
Isolated data have limited value. Inter-connected data can express much more complicated information. HTML is a perfect protocol to describe relationships, if not link could be unstable. The connections of information within a database is still of limited scope.
Many data centers are required to process data of billions of users in regular internet. A large data center has capacity for tens of thousands of servers. Every task is processed by groups of servers using parallel algorithm. A army of maintenance personals are required for daily operation.
No data center is required when storing personal data by individual. However, some kind of container, a light-weight clouds OS, will be necessary, like Pod promoted by Solid platform. The container shall be low cost, lightweight, and easy to operate by layman. Compared with the Homomorphic Encryption that certain block-chain algorithm is using, Leither method is simpler and more efficient. One who owns the storage device owns the data and defines the rules of application.
File
File is a set of program and data stored on media. File is constructed according to certain format, usually referred to as data structure. Operation system manages files with File System in a tree-like data structure. File wraps the access of data on media and optimizes the procedure.
WWW (World Wide Web)
In a standard webpage, HTML5 is used to render content of webpage, CSS style and JavaScript tasks. A page can store and render its data relatively comprehensively. Through interconnections among pages, large scale and complicated information can be constructed.
Shortcomings:
Webpage only contains information of frontend, without backend data. Big Techs take monopolistic control of user data by owning the backend servers.
The links that connect webpages are unstable due to the fragility of server and volatility of its location.
Lack of definition of elements such as server, user, etc.
Gene
The basic unit that contains information of life. DNA or RNA stores a complete segment of hereditary material, which records the adaptation of life to its environment. Life copies hereditary information from generation to generation, while adapting to new environment by variation.
Limitation: Every iteration takes a generation, with restrictions on the amount of information inherited and weak ability to deal with complicated changes.
MEME
Richard Dawkins introduced the concept of MEME in his original book, the Selfish Gene. Similar to gene, MEME is defined as the base unit that contains information of a civilization. MEME can be used to record all the knowledge of human society. It can be remembered in human brain, propagated among people, or stored in any kind of media.
Human brain uses pattern recognition to model the world. During its interaction with the environment, every reflection by the mind is an iteration of evolution. Compared with gene, MEME can evolute faster than gene by many orders of magnitude.
Origin of MiMei
Entropy is one of the most important measurements in the universe and human society. The 2nd law of thermodynamics points out the direction along which the universe evolves. That is why entropy is also called Arrow of Time. Erwin Schrodinger in his book, What is Life, declares that what an organism feeds upon is negative entropy.
Life stores information of its environment in gene. The heredity keeps the adaptation and the variation deals with the changes of environment. Life simply executes the instructions locked in its gene. Once human developed self-conscious, human brain begins to simulate adjustment made to its environment and process the information. The basic unit of such information is called Meme by Richard Dawkins.
After the birth of internet, the method for processing information has changed and a new unit of information is required. We call it MiMei. Similar to Gene and Meme, MiMei contains a piece of information and the rules applicable to it.
Implementation of MiMei
Based on Sec 2.7 Summary of Design, the following functionalities have be implemented.
- MiMei operation: create, manage, save, render and send
- MiMei creation by content's topic, each MiMei with a unique ID
- Right to set permissions on MiMei by user
- Referential relationship among MiMei
- MiMei can run or be saved on multiple nodes
- Version of MiMei, similar to Git, to solve problem of data consistency
- Synchronization among nodes in real-time or after backup
- Type of database and file, stream type coming soon
- File system as a special system type, based on database and file
- Independent space in database and file system for every MiMei
- Heredity and evolution of MiMei by continuously backing up changes
- User can also facilitate variation by fork Leither is implemented as MiMei container system, and also a decentralized cloud OS. MiMei API
MiMei Features
| Feature | Detail |
|---|---|
| Unique label | MiMei ID with version number can describe a piece of information consistently and precisely |
| Complex information rendering | Support database, file system, referential relationship and App. Describe complicated information |
| Free Data Migration | Construct relationship chain or group. MiMei can flow among different nodes according to user behaviors |
| Multi-dimension search | User can add tags to MiMei and customize multi-dimension search result through multiple nodes |
The above features will be explained in detail later.
Traditional method
Path of a resource in regular protocols,
WWW: http://server_ip[:port#]/path/file_name[parameter=value]
OS: disk_partition/path/file_name
In all of the above methods, every thing is indecisive, server IP, port number, path, file name, parameter. Existing protocols cannot describe location of a resource precisely. In IPFS, a unique ID is generated by hashing the content of a file, however a new ID must be created if any change happens to the file. This kind of uncertainty is troublesome to normal operations and referential integrity.
In MiMei system there is a brand new solution: MiMei ID
In the decentralized network of Leither, all resources are described by ID, including user, node, content, application, etc. ID of user and node is generated by encryption key pairs. Label ID of file or data block is based on synopsis of its content.
There are two sets of labels for external and internal reference.
- External label is MiMei ID
MiMei ID is the unique ID of a MiMei object, never changes after its creation.
MiMei ID can be used in index, operation and reference of the object.
MiMei data is stored in file or database. Each time the data is changed, a new version of backup is created with an ID based on the synopsis of its content. The version number increments from 0.
For convenience the latest version of backup data is called last. - Internal label is content synopsis ID
Every backup version of MiMei is pointed to by a file or database ID, which is generated by hashing synopsis of content.
MiMei Information
Information of MiMei can be defined as:
type MiMeiInfo struct {
Author string //creator
AppType string //associated application type
Ext string //extension
Mark string //MiMei ID
Create time.Time //timestamp of creation
Right uint64 //authorized rights
}MiMei ID unchangeable
MiMei ID is based on information of creator, associated application, MiMei type and MiMei label. Once created, MiMei ID never changes no matter how its content changes.
Access MiMei data with MiMei ID and version number
New version is created while MiMei being edited or backed up. New version ID is based on synopsis of its content, which is read-only after backup. Historical MiMei data can be retrieved by MiMei ID combined with version number.
MiMei can support data storage of most internet applications with its support of file system and database.
MiMei granulates information. It is recommended to redefine the following types of information with MiMei data type, aka Mimeimization:
- Content needed to be indexed
- Content for sharing among users
- Content that migrate among nodes
Mimeimization can be executed beforehand, or on demand. The latter is actually a split. A new MiMei object detached from the original one. A referential relationship keeps the tie.
Leither supports database and file system. User can use traditional development method if only its own data is concerned. In this case, one stand alone MiMei object is constructed for the user or application.
Both MiMei file system and database support version mechanism.
- cur: the current working copy, not yet backed up.
- last: the latest backup copy, represents the newest confirmed content
- release: tested and ready for release
MiMei ID and cur for the newest data
MiMei ID and last for the latest backup
MiMei ID and release for the most stable data
MiMei ID and version number for any historical data
Isolated data cannot describe complicated information. With unique label of MiMei, it is possible to establish relatively stable relationship structure. Information can be saved in MiMei file or database, with format defined by its associated application.
Association between MiMei is established by reference. Referential information includes MiMei ID and number of references. Ordinarily the semantic association of data is interpreted by its associated application. Leither system cannot access App data, so it is the App's job to maintain the reference of MiMei by calling corresponding API.
MiMei application can generate information of references according to semantics.
API: MMAddRef MMDelRef MMGetRef
Most MiMei contains only granulated piece of information, the whole picture can be described through association of MiMei.
Originally file system was designed for mainframe where the number of applications and volume of data is limited. File system increased efficiency by saving applications from the task of managing and accessing storage media. With the development computer and internet, especially mobile internet, the number of users, Apps and data volume all increased explosively. The following shortcomings of file system began to appear.
- File name cannot precisely label a file
- Insufficient index information, only path and file information.
- One file might have multiple duplicated copies
In Leither, traditional file system is converted into MiMei framework with the following improvements,
- Unique label: Every data file has a unique ID generated from its synopsis, as its label.
- cur version: Current version as target of frequent data access.
- Reference count: Do not copy data, increase count of reference instead. Use unique label to access its data.
The directory structure of file system is also granulated (Ch6.1). Directory that is indexed or shared with high hit count shall be Mimeimized. Currently Leither manages file directory with JSON file, which performs poorly when the number of files is too big. In future database will replace JSON. The algorithm for generating synopsis is similar to that of Merkle tree.
- MiMei Current File
The current version of MiMei file, both readable and writable. - MiMei Version File
The read-only backup file of MiMei, created by backing up MiMei current file. System assigns a version number to each version of backup file. - Mac file
Read-only file labelled by Mac ID of the file's content. Mac file can be referred to by MiMei object. MiMei version file is a Mac file referred to by a version number. Temp file can also be convert into Mac file. - Temp file
Created by MFOpenTempFile. After data is written into it, converted to MAC file by MFTemp2MacFile. - MiMei File System
A built-in special MiMei type of Leither system. It can be created by MMCreate, or placed in webdav directory, or created by Leither automatically using configure file with .mmfs extension, or opened by MFOpenByPath as file system object. - Operation System File and Directory
Create a link to a directory or file in webdav, the linked directory or file can be opened by MFOpenByPth. - MiMei Root Directory
webdav is the general entrance to access MiMei on a node. File or Directory can be linked into webdav, or a configure file can be create that points to a MiMei object on the node.
- MiMei File
MMCreate: Create MiMei file, take a api.MM_File as input, return MiMei ID.
MMOpen: Open MiMei file, return handle ID for file operation. If ver is "cur", file is writable, otherwise read-only. - Open File by Path
MFOpenByPath: Open file in MiMei file system or Leither file system. - Open Mac File
MFOpenMacFile: Open a Mac file under a MiMei. - Open Temp File
MFOpenTempFile: Open a temporary file for read or write. MFTemp2MacFile: Convert to Mac file after read or write. - Close File
All the files need to be closed after operation, except temporary file. Considering the inevitable off-line scenarios, all handles of file operation will be closed after timeout.
- Object Operation
MFSetObject MFGetObject - Byte Array Operation
MFSetData MFGetData - Status Query
MFGetSize MFStat MfIsExist MFGetMimeType - Directory Operation MFReadDir
- File System Operation
FSFind FSMkDir FSRemoveAll FSStat FSRename - Others
MFTruncate MFCopy
There are two underlying databases. One is based on LevelDB and the other BoltDB. Both are revised in underlying code. LevelDB is used for current version to support read and write operations. Its consistency is based on time sequence. During a write transaction, changed data will be checked to see if it is modified by any other party during the transaction. If it is, transaction fails and initial caller will be required to redo the transaction. BoltDB is used for backup version and read-only.
Leither API refers to Redis, support five types of data structure: string operation, hash, list, set, ordered set, and transaction.
- Transaction Begin Commit Rollback
- String Set Get Incr IncrBy Strlen
- Hash Hmclear Hdel Hlen Hget Hmget Hmset Hgetall Hkeys Hscan Hrevscan HincrBy
- List Lpush Lpop Rpush Rpop Lrange Lclear Lmclear Lindex Llen Lset
- Set Sadd Scard Sclear Sdiff Sinter Smclear Smembers Srem Sunion Scan
- Odered Set Zadd Zcard Zcount Zrem Zscore Zrank Zrange Zrangebyscore Zremrangebyscore Zrevrange Zrevrangebyscore Zmclear Zclear ZincrBy
Leither API supports 40+ development languages, particular HTML5 because large number of frequently used applications are developed with HTML5, including website, App, Applet, etc. HTML5 is not perfect for the lack of standards for backend data processing and business logic. Leither makes up the missing functionalities of HTML5 in building cloud applications, including user authentication, application system, cloud file system and database, decentralized domain name resolution, data redundancy and load balance.
Business Logic in Front-end
By default business logics shall be executed in the front-end. The benefit is that Leither application development will be similar to single page application development, simple and easy.
Thin Node
PC and mobile phone are not suitable for long term stable service. Server is maintenance heavy and expensive. Router and TV box are usually proprietary system with restricted accessability. NAS and other customizable hardware are preferred devices. Less powerful CPU is ideal to support fine granulated Leither applications.
Special Optimization
Because Leither node device is intended to be maintained by layman, node setup and application installation must be dummy proof. Node is optimized for home network, with the unique decentralized domain name resolution of Leither system, Leither can provide service similar to commercial cloud server.
Leither node can also be optimized to support batch execution and back-end tasks.
Mimeimization of Application
Leither provides comprehensive application system, within which an application is actually a type of MiMei. Refer to Application System Documentary
Most frequently used applications are developed in HTML5, including website, App, applet, etc.
HTML5 provides only localized data storage (Web Storage and IndexedDB), without setting up the standard for backend data and business logic. The result is HTML5 applications depend heavily on the backend to process business logic and data. Front-end developers fall to the lowest stratum among developers for not handling business logic and core data.
Leither provides functions of authentication and authorization, application system, cloud file system and database, decentralized domain name resolution, error tolerance and load balance.
Besides capable of building complete cloud applications, all of the applications and data within Leither system is of MiMei type, which can freely migrate among nodes. MiMei also enables load balance mechanism of Leither. All those powerful features make Leither ideal for building cloud service platform. The best practice of Leither App development is to execute business logic in the frontend, therefore greatly reduce the load on the backend. The same server running Leither can outperform traditional platform 100 times or more.
Knowledge of HTML5 is sufficient for most of the work for developing cloud services or website, App and applet independently. The workflow is almost identical to a HTML single page application.
Leither is developed with weak CPU and small memory devices in mind. Leither applications consume very little system resources of the node, because data and business logic are executed in the frontend, usually user's browser or App.
For traffic heavy applications, traffic load can be efficiently balanced by nodes in users' home network, which is cheap. The cost of traffic is only 1/10th of regular web service.
Both Leither system applications and data support docker-like sandbox. Both file and database support versioned backup. One physical machine can support multiple users, multiple applications and their datum. Application and data can flow among different physical machines freely without changing domain name or link address.
Elastic service of Leither platform is enabled by decentralized domain name and MiMei mechanism.
The meaning of life is the spread of its information
Gene and Meme are the carriers of information of life, whose purpose is to occupy time and space as much as possible. Similarly the purpose of MiMei is the propagation of meaningful information.
Hot data is more valuable
FileCoin conveyed a misunderstanding that data is asset. Cold (static) data is actually a debt, for device, storage and network all cost money. The value of data appreciates only when data turn into traffic (get hot). Services to support data is mainly the operations of save and retrieval. However, services to support data traffic involve the creation or collection of content, save, propagation, render, value realization and business development. Each time a content is displayed, more value is created. The number of renders and the value of each display determine the value created by the content.
The flow of information is a business requirement
In traditional network, data backup, error tolerance, load balance and elastic cloud services are all migration of data in essence.
Information can flow in two different situations:
- Save voluntarily.
Usually when the MiMei data is highly valuable, it can be saved for future benefit. - Save on demand.
When a node is in danger of overloading, it may ask other nodes to help in backup, error tolerance, load balance, etc.
Procedure of save:
- Mimeimize information
After information is MiMeimized, shared information can be backed up as file block. - Duplicate data over node network
With proper authorization, copy MiMei block from one node to another - Update routing information of MiMei
Terminal content consumer will get routing information of MiMei first, then accessing its data. Routing information includes node information and the newest changes of MiMei. After routing information is updated, user can get updated MiMei information.
Ch2.6 mentioned four parts of social media model: Me, Contact, Application and Message. In Leither the corresponding model is defined as below:
| Category | Detail |
|---|---|
| Me | Authentication and authorization of user |
| App and service | App and service of the user |
| Message | Manage interactions between user and Apps |
During the exchange of information and services between user(node) and user(node), there must be a mechanism of quantitative settlement. Currency is the tool to settle business in daily life. In Leither network it is the credit that can be exchanged between users.
Every node can grant other nodes credit, which authorize the others to use its services. The difference between services provided and received is the equivalent of a IOU that a borrower gives the lender, or an cryptocurrency issued by the borrower. Cryptocurrency issued by a group is actually a certificate of contribution made by its members. The contribution could be anything valuable to the group, such as resource, money or work done for the group. This certificate can be used to exchange for group services or other value from its members.
Object of credit
In reality fiat currency is the object of the highest credit rating. In Leither network, services exchanged between nodes, such as bandwidth, storage and CPU, can serve as the object. Their order of importance is bandwidth > storage > CPU. Because of the characteristics of node network, bandwidth is the preferred base unit of settlement between nodes. All the other forms of services can be converted into bandwidth with a certain coefficiency. Fiat currency can also be used as object of credit.
Individual credit model
is a type of User-to-User (P2P) L/C or IOU.
- Letter of Credit (L/C) L/C can be understood as the amount of services allowed to use before payment, whereas IOU is the amount of services that have been used without payment. A L/C includes information of the authorizer, the authorized, unit of credit, maximum credit, signature of authorizer.
- IOU and Currency
An IOU includes information about lender, borrower, unit of credit, amount, signature of borrower. For borrower, IOU is equivalent to issuing Cryptocurrency.
How to buildup credit
- Initial grant of credit
If there were a relationship between nodes, there should have been certain basis for mutual trust, so that certain amount of credit could be granted according to the magnitude of that relationship. - Periodical credit increase
After a relationship between nodes established, credit can be increased periodically. - Mutual help
Each node is capable of providing useful services. Credit can be acquired by proactively providing valuable service to the others, including data backup, load balance, routing, search, or mediator service. - Credit revocation
User can revoke credit granted to others in case of misbehavior. - Acquisition of credit
User could gain extra credit through off-line purchase or negotiation. - Credit exchange
User can exchange for more credit through 3rd party. Reference to Organization and Consensus.
Trading procedure
With sufficient credit, user can exchange for services with credit, afterward a record will be added into public ledger. When user exceeds its credit cap, service will stop.
Agency Service
When there is no direct connection between nodes, an agent node could be used to mediate the credit exchange. It is also possible to use other trusted media to facilitate the exchange, such as fiat currency, cryptocurrency or group currency mentioned in 2nd part.