DE102018009320A1 - Working methods of computer systems in a blockchain computer network to protect the integrity of a distributed database - Google Patents

Abstract

Working methods of computer systems (10) in a blockchain computer network (60) to protect the integrity of a distributed database by means of self-regulating network formation (190), communication control (320) and finding consensus (170) between the computer systems (10) (see Figure 1). The working method solves the problem of various centralization risks and attack vectors that exist in various forms in classic blockchain computer networks (60). The working process enables the creation of an open multi-agent system (180) to solve these problems. The integrity of important distributed databases (e.g. financial data, cryptocurrencies, etc.) and their secure modification is, among other things, This ensures that the balance of power in finding consensus (170) is determined not by the environmentally harmful performance of the individual computer systems (10) but by an organically built system of trust relationships among the computer systems (10). The core of the solution relates to the calculation and exchange of ratings (80) of computer systems (10), the evaluation of transaction messages (40), a firmly defined dynamic of negotiations between the computer systems (10), but also two random values, which various predictions and others Make it impossible for attackers to prepare. The rating (80) plays a defined, important role in all procedural steps

Description

Technical field

One or more implementations relate to generally distributed blockchain databases, but in particular open multi-agent systems, the networking of computer systems in a blockchain computer network and the automatic finding of consensus for transaction messages between these computer systems.

State of the art

1. 1. Komplett vertrauensfreie Blockchain-Computernetzwerke (60), in denen kein Computersystem (10) einem anderen vertraut und in welchen man Regeln bzw. die richtige Konsensfindung (170) durch ein System von finanzieller Sanktion und Entlohnung durchsetzt.
2. 2. Teilweise zentralisierte Blockchain-Computernetzwerke (60), in denen wenige vordefinierte Computersysteme (10) komplettes Vertrauen des ganzen Netzwerkes genießen und immer das letzte Wort in Konfliktsituationen bzw. bei der Konsensfindung (170) haben. Sie regeln aud diese Weise die Arbeitsweise des gesamten Blockchain-Computernetzwerks (60).

Blockchain technologies have established themselves as procedures for securing the integrity of distributed databases and their change transactions ( 30th ). These databases are independent of central instances of trust (e.g. authorities, audit departments, central servers, etc.) and therefore have various advantages in business, particularly with regard to process optimization, cost reduction and protection against manipulation / fraud. Automation of trust through technology is tempting, but various risks quickly arise through centralization. There are currently two rough approaches:

1. 1. Completely trust-free blockchain computer networks ( 60 ) in which no computer system ( 10th ) trust another and in which one rules or the right consensus finding ( 170 ) enforced by a system of financial sanction and remuneration.
2. 2. Partially centralized blockchain computer networks ( 60 ), in which few predefined computer systems ( 10th ) enjoy complete trust of the whole network and always have the last word in conflict situations or in finding consensus ( 170 ) to have. In this way, they regulate the functioning of the entire blockchain computer network ( 60 ).

In the second case, one can no longer speak of a purely decentralized system and therefore cannot enjoy all the advantages: there are obvious points of attack and inefficiencies.

In the first case, irreversible inequality is also achieved through the disproportionate nature of financial remuneration (e.g. through cryptocurrencies such as Bitcoin) and the sanctions (using a "proof-of-work" (220) CPU load). This reward stimulates the rapid growth of powerful, environmentally hostile computer systems (see "bitcoin miners"), which have a disproportionate influence on the consensus building in the entire blockchain computer network ( 60 ) win. In this case, a kind of centralization arises indirectly through the use of powerful computer systems. The balance of power in finding consensus ( 170 ) therefore shift in the direction of financial strength: there are obvious points of attack or risks.

Centralization is catastrophic for distributed databases, but the risks are mostly underestimated. Centralized systems offer a target or target for hackers, corruption, abuse of power, technical and human error, etc. Only by avoiding a shift in the balance of power in the consensus building of a blockchain computer network ( 60 ) you can minimize these risks sustainably.

In classic blockchain technologies, performance also plays a role in attacks by malicious computer systems ( 250 ) and especially in the case of SPAM attacks ( 260 ) a major role. Especially if the network topography ( 200 ) is known, attackers can use individual parts or even the entirety of the blockchain computer network ( 60 ) paralyze and / or split up and try new facts in the distributed blockchain database ( 20th ) to accomplish. These attacks are partially prevented by, among other things, the achievement of the "Proof-of-Work" (220), also known as PoW. At the PoW, the attacker would have to 40 ) Invest CPU resources and solve a cryptographic puzzle. However, with enough resources, this is a plausible attack vector on a blockchain computer network ( 60 ).

The challenge remains an efficient, secure, and completely decentralized blockchain computer network ( 60 ) to ensure that the integrity of a distributed database is ensured without the risk of centralization through a shift in power using resources.

When it comes to performing distributed tasks, in theory systems of distributed artificial intelligence or multi-agent systems (MAS) ( 180 ), are used. MAS are already used in solutions in other areas, e.g. in the execution of logistical tasks using swarm robotics, in distributed energy systems, in the analysis of big data, etc.

To achieve their goals, agents must be in an open MAS ( 180 ) engage in social interaction, such as cooperation, coordination, negotiation and conflict resolution (see Wooldridge 2009). Both the dimension of trust and the dimension of the associated interactions such as coalition, argumentation and recommendation (see J. Granatyr 2015 p.2) play a role in these activities. Agents can treat their trust as a numerical value that expresses their opinion how trustworthy they consider another agent to be (cf. Sabater and Sierra 2005, pp. 33-60). On the other hand, recommendations are collections of opinions received by other agents (see Nunes 2011)

When working through over 250 known academic articles on MAS ( 180 ), it is noticeable that only 110 deal with trust and reputation systems. Most implementations are closed MAS ( 180 ). For certain cases, however, solutions for open MAS ( 180 ) who also deal with fraud assumptions. There are currently around 15 scientific papers. None of these deal with topics such as risk management, hard security (eg using asymmetric keys) or sources of information for the respective computer systems ( 10th ) in the MAS ( 180 ) - these are, however, issues that one in a blockchain network environment ( 60 ) would need.

Underlying task

1. 1. die Netzwerkformierung (190), die Kommunikationssteuerung (320) und die Konsensfindung (170) zwischen den Computersystemen (10) dynamisch und selbstregulierend stattfindet; und dabei
2. 2. die Machtverhältnisse bei der Konsensfindung (170) nicht durch die Leistungsstärke der einzelnen Computersysteme (10) sondern durch ein organisch aufgebautes System von Vertrauensbeziehungen unter den Computersystemen (10) bestimmt werden; und dabei
3. 3. eine autonome feingranulare PoW-Steuerung (220/240) die Datenüberflutung von einzelnen Computersystemen (10) verhindert; und dabei
4. 4. eine sichere/deterministische Informationsprozessierung und - speicherung von verteilten Datenbeständen in einer Blockchain-Datenbank (20), auch in einem bösartigen Netzwerkumfeld und unter einer umfangreichen und vielfältigen Angriffsbelastung (300), sichergestellt ist.

The invention specified in claim 1 is based on the problem for the computer systems ( 10th ) in a completely decentralized blockchain computer network ( 60 ) to define a working procedure in which an open multi-agent system ( 180 ) arises, and thereby:

1. 1. the network formation ( 190 ), the communication control ( 320 ) and finding consensus ( 170 ) between the computer systems ( 10th ) takes place dynamically and self-regulating; And thereby
2. 2. the balance of power in finding consensus ( 170 ) not by the performance of the individual computer systems ( 10th ) but through an organically built system of trust relationships among the computer systems ( 10th ) be determined; And thereby
3. 3. an autonomous fine-granular PoW control (220/240) the data overload of individual computer systems ( 10th ) prevents; And thereby
4. 4. Secure / deterministic information processing and storage of distributed data in a blockchain database ( 20th ), even in a malicious network environment and under an extensive and diverse attack load ( 300 ) is ensured.

The solution and its beneficial effects

The underlying object is achieved by the features listed in claim 1.

The solution is therefore a firmly defined working method of a computer system ( 10th ) in a blockchain computer network ( 60 ). This working process consists of three self-regulating main work steps, the network formation ( 190 ), the communication control ( 320 ) and finding consensus ( 170 ), which are divided into several subtasks.

By continually re-adjusting various factors in the working process of each computer system ( 10th ), a process of distributed machine learning arises, an open multi-agent system ( 180 ), at the level of the entire blockchain computer network ( 60 ). Over time, emerge within the blockchain computer network ( 60 ) strong relationships between different computer systems ( 10th ) where a high level of mutual trust / rating ( 80 ) prevails. As a result, these computer systems ( 10th ) the connection probability is always higher and the duration of the connections and the acceptance of the exchanged transaction messages ( 40 ) grows. A kind of virtual backbone arises from computer systems ( 10th ) who have been around for a long time and have a high level of trust / rating ( 80 ) have built up. This rating ( 80 ) a computer system ( 10th ) lose quickly if it doesn't follow the rules. The rating ( 80 ), however, can be built up only with difficulty / slowly. It is a group dynamic that is known from sociology and that is artificially created here.

• • das gesamte Blockchain-Computernetzwerk (60) sich effektiv gegen verschiedene Angriffsszenarien und fehlerhafte Computersysteme (10) schützen und sich regenerieren kann. Tests zeigen, dass die Konsensfindung (170), auch bei einem 90%-Anteil von bösartigen Computersystemen (250) im Blockchain-Computernetzwerk (60), nicht gebrochen werden kann,
• • im Rahmen der Konsensfindung (170) gültige Transaktionsnachrichten (40) schnell und unter einer relativ geringen Netzwerkbelastung von Computersystemen (10) angenommen werden,
• • im Rahmen der Konsensfindung (170) ungültige Transaktionsnachrichten (40) schnell und unter einer relativ geringer Netzwerkbelastung von Computersystemen (10) verworfen werden,
• • der Einsatz von Ressourcen und leistungsstarken Computersystemen (10) keine Auswirkung auf die Machtverschiebung bei der Konsensfindung (170) hat, insbesondere je länger das Blockchain-Computernetzwerk (60) bereits existiert,
• • der Einsatz von Ressourcen und leistungsstarken Computersystemen (10) nahezu keinen Vorteil bei den Angriffen durch bösartige Computersysteme (250) bietet,
• • die Konsensfindung (170) in einem dynamischen, organisch gebildeten Vertrauenssystem unter den Computersystemen (10), effizienter und sicherer zu einer Konsensfindung (170) führt,
• • die Integrität der Blockchain-Datenbank (20) gesichert wird und nicht durch verschiedene Zentralisierungsrisiken untergraben werden kann.

The result of the working process is that:

• • the entire blockchain computer network ( 60 ) effectively against various attack scenarios and faulty computer systems ( 10th ) can protect and regenerate. Tests show that consensus building ( 170 ), even with a 90% share of malicious computer systems ( 250 ) in the blockchain computer network ( 60 ), cannot be broken,
• • as part of building consensus ( 170 ) valid transaction messages ( 40 ) quickly and under a relatively low network load on computer systems ( 10th ) are assumed
• • as part of building consensus ( 170 ) invalid transaction messages ( 40 ) quickly and under a relatively low network load on computer systems ( 10th ) are discarded,
• • the use of resources and powerful computer systems ( 10th ) no impact on the shift in power in finding consensus ( 170 ), especially the longer the blockchain computer network ( 60 ) already exists,
• • the use of resources and powerful computer systems ( 10th ) almost no advantage in attacks by malicious computer systems ( 250 ) offers,
• • finding consensus ( 170 ) in a dynamic, organically formed trust system among the computer systems ( 10th ), more efficient and safer to find a consensus ( 170 ) leads,
• • the integrity of the blockchain database ( 20th ) is secured and cannot be undermined by various centralization risks.

Embodiment and explanation

1. 1. Netzwerkformierung (190)
2. 2. Kommunikationssteuerung (320)
3. 3. Konsensfindung (170)

How 1 shows, the computer systems follow ( 10th ) in the blockchain computer network ( 60 ) a working process consisting of three main steps:

1. 1. Network formation ( 190 )
2. 2. Communication control ( 320 )
3. 3. Finding consensus ( 170 )

• • im internen Zustand,
• • im Netzwerkumfeld und
• • in der Kommunikation

des Computersystems (10) an.Each of these steps consists of several subtasks. Each of the work steps continuously adapts to the current situation through learning processes:

• • in internal condition,
• • in the network environment and
• • in communication

of the computer system ( 10th ) at.

1. 1. Bewertung (3a und 3b): Festlegung und/oder Anpassung von mehreren relevanten Faktoren zu den bekannten Computersystemen (10)
2. 2. Verbindungsaufbau (3a und 3b): Eingehende (100) und ausgehende (110) Verbindungen verwalten
3. 3. Vorstellung (4): Austausch der Gesamtheit oder einer Teilmenge der Kontaktliste (160) mit Nachbarn (70)
4. 4. Meinungsaustausch (5): Kommunizieren des eigenen Ratings (80) über einen bestimmten Nachbarn D (70) an den Nachbarn A (70), auf Anfrage von A

The main step of "network formation" (190) (see 3a , 3b , 4th and 5 ) consists of four continuous subtasks, in which the arrangement is immaterial, since these can possibly be carried out asynchronously:

1. 1. Evaluation ( 3a and 3b ): Definition and / or adaptation of several relevant factors to the known computer systems ( 10th )
2. 2. Establishing a connection ( 3a and 3b ): Incoming ( 100 ) and manage outgoing (110) connections
3. 3rd performance ( 4th ): Exchange all or part of the contact list ( 160 ) with neighbors ( 70 )
4. 4. Exchange of views ( 5 ): Communicate your own rating ( 80 ) via a certain neighbor D (70) to neighbor A (70), at the request of A

1. 1. Rating $\text{'}{R}_b^a\text{'}$
   
   : das Vertrauen (80), welches das Computersystem (10) A zu dem Nachbarn (70) B hat
2. 2. „Neugier“ ‚Slots_n‘: Anzahl der verfügbarer offenen Slots (120), die exklusiv für eingehende Verbindungen (100) mit neuen/unbekannten Computersystemen reserviert sind
3. 3. Gruppenzugehörigkeit (130) ‚Groups_a, Groups_b‘: gemeinsame Gruppenzugehörigkeit (130) zwischen dem Computersystem (10) A und dem Nachbarn B (70)
4. 4. Verbindungsqualität ‚Connection_q‘: durchschnittliche Latenz der Datenverbindung zwischen dem Computersystem (10) A und dem Computersystem (10) B
5. 5. Persönlichkeit (140) ‚Personality_a, Persanality_b‘: ein kontinuierlich neu berechneter Zufallswert in jedem Computersystem (10) eines Blockchain-Computernetzwerks (60), durch den fortlaufend eine zufällige Netzwerktopografie (200) entsteht und dadurch Angriffsszenarien wie „Eclipse“ (270) unmöglich gemacht werden

In the subtask “Assessment”, the computer system ( 10th ) A multiple factors to a new or known neighboring computer system ( 70 ) B determines:

1. 1st rating $\text{'}{R}_b^a\text{'}$
   
   : the trust ( 80 ) which the computer system ( 10th ) A to the neighbor ( 70 ) B has
2. 2. "Curiosity"'Slots _n ': number of available open slots ( 120 ) exclusively for incoming connections ( 100 ) are reserved with new / unknown computer systems
3. 3. Group membership ( 130 ) 'Groups _a , Groups _b ': shared group membership ( 130 ) between the computer system ( 10th ) A and the neighbor B (70)
4. 4. Connection _quality 'Connection _q ': average latency of the data connection between the computer system ( 10th ) A and the computer system ( 10th ) B
5. 5. Personality ( 140 ) 'Personality _a , Persanality _b ': a continuously recalculated random value in every computer system ( 10th ) a blockchain computer network ( 60 ), through which a random network topography ( 200 ) arises, making attack scenarios like "Eclipse" (270) impossible

(80) von Computersystem (10) A über ein anderes Computersystem (10) B. Das Rating wird durch die Gesamtheit oder eine Teilmenge von folgenden Kennzahlen ermittelt:

1. 1. Dauer der Beziehung ‚Connection_t‘: wie lange kennt das Computersystem (10) A das andere Computersystem (10) B ?
2. 2. Dauer der Kommunikation ‚Communication_t‘: wie lange besteht/bestand aktive direkte Kommunikation zwischen den beiden Computersystemen (10) ?
3. 3. Nachbarmeinung ‚Reputation‘: welches gewichtete Rating (80) geben die Nachbarn (70) von Computersystem (10) A über das Computersystem (10) B ? Dabei spielt das Rating (80) des jeweiligen Nachbarn (70) eine Rolle bei der Gewichtung des Gesamtratings der Nachbarn (70) über das Computersystem (10) B.
4. 4. Neuheitswert der Transaktionsnachrichten (40) ‚Transaction_n‘ : welcher Anteil von den Transaktionsnachrichten (40), die von Computersystem (10) B an Computersystem (10) A geschickt wurden, waren neu bzw. waren nicht bereits bekannt? Dieser Wert ermittelt sich aus den Ergebnissen des Hauptarbeitsschrittes „Konsensfindung“ (170).
5. 5. Gültigkeit der Transaktionsnachrichten (40) ‚Transaction_v‘: welcher Anteil von den Transaktionsnachrichten (40), die von Computersystem (10) B an Computersystem (10) A geschickt wurden, waren gültig bzw. mussten später nicht korrigiert oder verworfen werden ? Dieser Wert ermittelt sich aus den Ergebnissen des Hauptarbeitsschrittes „Konsensfindung“ (170).
6. 6. Änderung der Transaktionsnachrichten (40) ‚Transaction_c‘: wie oft ändert das Computersystem (10) B die Meinung über die Gültigkeit der einen oder der anderen Transaktionsnachricht (40) ? Dieser Wert ermittelt sich aus den Ergebnissen des Hauptarbeitsschrittes „Konsensfindung“ (170).

$$\begin{array}{c}\text{'}{R}_b^a=f(Connectio{n}_t,Communicatio{n}_t,Reputation,\\ Transactio{n}_n,Transactio{n}_v,Transactio{n}_c,\dots )\text{'}\end{array}$$

The most important factor in the rating is the rating $\text{'}{R}_b^a\text{'}$

(80) from computer system ( 10th ) A through another computer system ( 10th ) B. The rating is determined by all or a subset of the following key figures:

1. 1. Duration of the relationship 'Connection _t ': how long has the computer system known ( 10th ) A the other computer system ( 10th ) B?
2. 2. Duration of communication 'Communication _t ': how long is / was there active direct communication between the two computer systems ( 10th )?
3. 3. Neighboring opinion 'reputation': which weighted rating ( 80 ) give the neighbors ( 70 ) from computer system ( 10th ) A about the computer system ( 10th ) B? The rating ( 80 ) of the respective neighbor ( 70 ) a role in the weighting of the overall rating of the neighbors ( 70 ) via the computer system ( 10th ) B.
4. 4. Novelty value of the transaction messages ( 40 ) 'Transaction _n ': which part of the transaction messages ( 40 ) by computer system ( 10th ) B to computer system ( 10th ) A were sent, were new or were not already known? This value is determined from the results of the main work step "finding consensus" (170).
5. 5. Validity of the transaction messages ( 40 ) 'Transaction _v ': which part of the transaction messages ( 40 ) by computer system ( 10th ) B to computer system ( 10th ) A sent were, were valid or did not later have to be corrected or rejected? This value is determined from the results of the main work step "finding consensus" (170).
6. 6. Change of transaction messages ( 40 ) 'Transaction _c ': how often does the computer system change ( 10th ) B the opinion about the validity of one or the other transaction message ( 40 )? This value is determined from the results of the main work step "finding consensus" (170).

$$\begin{array}{c}\text{'}{R}_b^a=f(\mathrm{C.}OnnectiO{n}_t,\mathrm{C.}OmmunicatiO{n}_t,ReputatiOn,\\ TransactiO{n}_n,TransactiO{n}_v,TransactiO{n}_c,...)\text{'}\end{array}$$

In the subtask "Establishing a connection" (see 3a and 3b ) of the main work step "network formation" (190) both outgoing (110) and incoming (100) connections are established. When establishing a connection, the identity of the computer systems ( 10th ) checked with the public keys of the respective key pair from public key encryption processes and a corresponding handshake (115). A computer system ( 10th ) each has a limited number of slots ( 120 ) (please refer 2nd ) for incoming connections ( 100 ) and a limited number of slots for outgoing connections ( 110 ). This separation is necessary to prevent both SPAM attacks ( 260 ) as well as Eclipse attacks ( 270 ) counteract. The number of slots ( 120 ) for incoming connections ( 100 ) exceeds the number of slots ( 120 ) for outgoing connections ( 110 ) (e.g. 9 and 7), see 2nd . As a result, the blockchain computer network ( 60 ) can grow more easily. A computer system ( 10th ) the goal is always to maintain at least 5-10 active connections.

Only through an active connection in a slot ( 120 ) are transactional messages ( 40 ) accepted and forwarded to the next main work step "Communication control" (320).

When deciding on an outgoing connection ( 110 ) 'Connectian _o ' with a neighbor ( 70 ) B, the following factors play a role in their entirety or as a subset: current connections, group membership ( 130 ), Rating ( 80 ), Connection quality, personalities ( 140 ) $$\begin{array}{c}\text{'}\mathrm{C.}OnnectiO{n}_O=f(\mathrm{C.}OnnectiOns,GrOup{s}_a,GrOup{s}_b,R_b^a,\\ \mathrm{C.}OnnectiO{n}_q,PersOnalit{y}_a,PersOnalit{y}_b...)\text{'}\end{array}$$

When deciding on an incoming connection ( 100 ) 'Connection _i ' with a neighbor ( 70 ) B the following factors play a role in their entirety or as a subset: current connections, group membership ( 130 ), Rating ( 80 ), Curiosity slots, personalities ( 140 ) $$\begin{array}{c}\text{'}\mathrm{C.}OnnectiO{n}_l=f(\mathrm{C.}OnnectiOns,GrOup{s}_a,GrOup{s}_b,R_b^a,\\ SlOt{s}_n,PersOnalit{y}_a,PersOnalit{y}_b,...)\text{'}\end{array}$$

In the subtask "Presentation" (see 4th ) of the main work step "network formation" (190) is sent by the computer system ( 10th ) A to the neighbors ( 70 ) B, on request from B, a part of 'Contactlist _b ∈ Contactlist' of the entire contact list ( 160 ) by A.

$$\text{'}Contactlis{t}_b=f\left(Contactlist,ShareMa{x}_b\right)\text{'}$$

The maximum number of communicated contacts 'ShareMax _b ' from computer system ( 10th ) A at the neighbors (B) is weighted by the following factors in their entirety or as a subset: Group membership ( 130 ), Rating ( 80 ), Connection quality $$\text{'}SHareMa{x}_b=f\left(GrOup{s}_a,GrOup{s}_b,R_b^a,\mathrm{C.}OnnectiO{n}_q,...\right)\text{'}$$

$$\text{'}\mathrm{C.}Ontactlis{t}_b=f\left(\mathrm{C.}Ontactlist,SHareMa{x}_b\right)\text{'}$$

The communicated selection is noted and reproduced in the same way on request. As soon as the factors calculated for B allow a better / larger maximum number of contacts 'ShareMax _b ', additional contacts will be revealed. As a result, computer systems ( 10th ) with higher rating ( 80 ) have a better chance of dealing with new / unknown computer systems ( 10th ) in the blockchain computer network ( 60 ) connect to.

In the subtask "exchange of views" (see 5 ) of the main work step "network formation" requires the computer system ( 10th ) A in its outgoing connection ( 110 ) to the neighbor ( 70 ) B, the rating determined by B ( 80 ) via the computer system ( 10th ) D. The result is weighted with the rating of A over B and flows into the entire neighboring opinion 'Reputation _d ' over D.

von A über D. $$\text{'}Reputatio{n}_d=f_2\left(f_1\left(R_d^b,R_b^a\right)+f_1(\dots )+f_1\left(R_d^n,R_n^a\right)\right)\text{'}$$

$$\text{'}{R}_d^a=f\left(\dots ,Reputatio{n}_d,\dots \right)\text{'}$$

This neighboring opinion later plays a role in determining the rating ( 80 ) $\text{'}{R}_d^a\text{'}$

from A to D. $$\text{'}ReputatiO{n}_d=f_{\mathrm{2nd}}\left(f_1\left(R_d^b,R_b^a\right)+f_1(...)+f_1\left(R_d^n,R_n^a\right)\right)\text{'}$$

$$\text{'}{R}_d^a=f\left(...,ReputatiO{n}_d,...\right)\text{'}$$

1. 1. Empfang: Transaktionsnachricht (40) empfangen und TbPoW-Anforderungen bestimmen
2. 2. TbPoW-Anforderung: Vom Sender der Transaktionsnachricht (40) eine Leistungsbestätigung, i.e. Nonce (210), anfordern
3. 3. Firewall: die Transaktionsnachricht (40) anhand des Nonce (210) aus TbPoW (240) akzeptieren oder verwerfen

The main work step "Communication control" (320) (see 6a and 6b ) monitors the incoming communication before it is made available to the next main work step "finding consensus" (170). It therefore fulfills the function of a firewall. The unique "Trust-bound-Proof-of-Work" (TbPoW) (240) is used, in which the nonce difficulty level ( 230 ) depending on the rating ( 80 ) of a neighbor ( 70 ) is determined. This main step consists of an incoming transaction message ( 40 ) from three subtasks:

1. 1.Receipt: transaction message ( 40 ) received and determine TbPoW requirements
2. 2. TbPoW request: From the sender of the transaction message ( 40 ) a confirmation of performance, ie Nonce ( 210 ), request
3. 3. Firewall: the transaction message ( 40 ) based on the nonce ( 210 ) from TbPoW (240) accept or reject

der durch folgende Faktoren, als Gesamtheit oder Teilmenge, bestimmt wird:

1. 1. Das aktuelle Rating (80) von B $\text{'}{R}_b^a\text{'},$
   
   höhere Werte mindern den Nonce-Schwierigkeitsgrad (230)
2. 2. Durchschnittliche Netzwerkbelastung ausgedrückt als durchschnittliche Anzahl der erhaltenen Transaktionsnachrichten (40) pro Nachbar pro Zeiteinheit ‚NetLoad‘
3. 3. Durchschnittliche Anzahl der Transaktionsnachrichten 840 erhalten von B pro Zeiteinheit ‚NetLoad^b‘, je größer dieser Wert im Vergeich zu der gesamten durchschnittlichen Netwerkbelastung , desto höher die Anforderung durch den Nonce-Schwierigkeitsgrad (230)
4. 4. Aktuelle CPU-Belastung vom Computersystem (10) A ‚CPU_a‘ : je höher die Belastung desto höher der Nonce-Schwierigkeitsgrad (230)

$$\text{'}TbPo{W}_b^a=f\left(R_b^a,NetLoad,NetLoa{d}^b,CP{U}_a,\dots \right)\text{'}$$

In the subtask “reception” of the main work step “communication control” (230) the computer system ( 10th ) A a new transaction message ( 40 ) from the computer system ( 10th ) B. Upon receipt, it determines a nonce difficulty level ( 230 ) $\text{'}TbPO{W}_b^a\text{'},$

which is determined by the following factors, as a whole or a subset:

1. 1. The current rating ( 80 ) from B $\text{'}{R}_b^a\text{'},$
   
   higher values reduce the nonce level of difficulty ( 230 )
2. 2. Average network load expressed as the average number of transaction messages received ( 40 ) per neighbor per time unit 'NetLoad'
3. 3. Average number of transaction messages 840 received from B per unit of time 'NetLoad ^b ', the larger this value compared to the total average network load, the higher the requirement due to the nonce difficulty level ( 230 )
4. 4. Current CPU load from the computer system ( 10th ) A 'CPU _a ': the higher the load, the higher the level of nonce ( 230 )

$$\text{'}TbPO{W}_b^a=f\left(R_b^a,NetLOad,NetLOa{d}^b,\mathrm{C.}P{U}_a,...\right)\text{'}$$

Transaction messages ( 40 ) that have already been received and processed with the same content from B are ignored / rejected and TbPoW (240) is then not requested. If the transaction message received ( 40 ) a reply to another transaction message sent from A to B ( 40 ), it is forwarded without TbPoW (240) directly to the main work step "finding consensus" (170).

von A an B kommuniziert. Jetzt muss B das TbPoW (240) unter Einsatz der eigenen CPU Leistungsstärke lösen, indem es den Nonce (210) bestimmt. Die Besonderheit hier ist, im Gegensatz zum klassischen PoW (220), wo die Nonce-Lösung (210) sich nur auf den Inhalt eines Datenpakets bzw. „Blocks“ bezieht, dass die Lösung von TbPoW (240) sich sowohl auf den Inhalt der Transaktionsnachricht (40) als auch auf den Sender und auf den Empfänger, und ggf. einer vom Empfänger festgelegten Zufallsvariable, beziehen muss. D.h. B muss das TbPoW (240) lösen, indem es nach dem Nonce (210) sucht, das sich zusammen in einem Datenpaket bestehend aus:

1. 1. der Transaktionsnachricht,
2. 2. einem Zufallswert vorgegeben von A,
3. 3. dem öffentlichen Schlüssel von B und
4. 4. dem öffentlichen Schlüssel von A

befindet.In the next subtask “TbPoW requirement” the calculated nonce difficulty level ( 230 ) $\text{'}TbPO{W}_b^a\text{'}$

communicated from A to B. Now B has to solve the TbPoW (240) using its own CPU performance by using the Nonce ( 210 ) certainly. The special feature here is, in contrast to the classic PoW (220), where the nonce solution ( 210 ) refers only to the content of a data packet or "block", that the solution from TbPoW (240) relates both to the content of the transaction message ( 40 ) as well as the sender and the receiver, and possibly a random variable defined by the receiver. Ie B must solve the TbPoW (240) by following the Nonce ( 210 ) that is composed of a data package consisting of:

1. 1. the transaction message,
2. 2. a random value given by A,
3. 3. the public key of B and
4. 4. the public key of A

located.

This means that every single transaction message ( 40 ) from a computer system ( 10th ) to someone else's own special nonce solution ( 210 ). Of course here are the levels of difficulty ( 230 ) much lower than in the classic PoW (220) approaches.

In the next subtask "Firewall", the computer system decides ( 10th ) A whether it is the transaction message ( 70 ) from B if no or wrong nonce solution ( 210 ) is obtained from TbPoW (240). The transaction message ( 40 ) are forwarded to the next main work step "finding consensus" (170).

1. 1. Empfang (7a): eine Transaktionsnachricht (10) erhalten und bewerten
2. 2. Antwort (7b): die Transaktionsnachricht (10) bestätigen oder ihr widersprechen
3. 3. Weiterleitung (7b): die Transaktionsnachricht (10) weiterleiten
4. 4. Annahme (8): die Transaktionsnachricht (10) als gültig und final annehmen

The main work step "finding consensus" (170) (see 7a , 7b and 8th ) has the primary goal that all computer systems ( 10th ) of the blockchain computer network ( 60 ) from a set of possibly contradicting transaction messages ( 310 ) only the same and valid transaction messages ( 40 ) and the related change transaction (s) ( 30th ) sustainably in the blockchain database ( 20th ) to install or chain. This main work step consists of four subtasks:

1. 1st reception ( 7a ): a transaction message ( 10th ) received and evaluated
2. 2nd answer ( 7b ): the transaction message ( 10th ) confirm or contradict it
3. 3. Forwarding ( 7b ): the transaction message ( 10th ) hand off
4. 4. Acceptance ( 8th ): the transaction message ( 10th ) as valid and final

das im Hauptarbeitsschritt „Netzwerkformierung“ (190) fortlaufend ermittelt wird. Sollte diese Transaktionsnachricht (40) A bereits bekannt sein, weil sie genau so von anderen Nachbarn (70) bereits empfangen wurde, z.B. von Nachbar (70) C und D, so erhält diese Transaktionsnachricht (40) ein Nachrichtenrating (90), das anhand von allen Ratings (80) der Nachbarn (70) errechnet wird, die diese Transaktionsnachricht (40) an A geschickt haben. $$\text{'}MessageRatin{g}_x=f\left(R_b^a,R_c^a,R_d^a,\dots \right)\text{'}$$

In the subtask “reception” of the main work step “finding consensus” (170), the computer system receives ( 10th ) A a transaction message X from the neighbor ( 70 ) B after it successfully firewall ( 320 ) has overcome. This transaction message ( 40 ) is rated by A: it gets the same news rating ( 90 ) like the rating of the neighbor ( 70 ) B: 'MessageRating _x = $R_b^a\text{'},$

which is continuously determined in the main work step "network formation" (190). If this transaction message ( 40 ) A already known because it is exactly the same from other neighbors ( 70 ) has already been received, e.g. by neighbor ( 70 ) C and D, this transaction message receives ( 40 ) a news rating ( 90 ) based on all ratings ( 80 ) the neighbour ( 70 ) is calculated that this transaction message ( 40 ) sent to A. $$\text{'}MessaGeRatin{G}_x=f\left(R_b^a,R_c^a,R_d^a,...\right)\text{'}$$

Dieser Abzug könnte z.B. der Median aus allen Nachbarratings (80) von A sein.If B previously another, the transaction message ( 40 ) X contradicting, transaction message ( 310 ) Z sent to A, it means that B changed his mind and now considers X valid instead of Z. In the event of a change of opinion, the news rating ( 90 ) of the new transaction message ( 40 ) X punished by getting a deduction: $\text{'}MessaGeRatin{G}_x=R_b^a-m\text{'}.$

This deduction could, for example, be the median of all neighboring ratings ( 80 ) of A.

From a lot of contradicting transaction messages ( 310 ) keeps the computer system ( 10th ) A always the transaction message ( 40 ) valid for the highest news rating at the time ( 90 ) and the factual state of the blockchain database ( 20th ) does not contradict A. A transaction message ( 40 ) is always discarded if A recognizes defects in the facts or in the form, e.g. missing or invalid signatures or addresses, "Double Spending" (290), etc.

1. 1. Neuheitswert ^lTransaction_n',
2. 2. Gültigkeit ‚Transaction_v‘ , und
3. 3. Änderung ‚Transaction_c‘

der von B erhaltenen Transaktionsnachrichten (40).When evaluating the transaction messages received ( 40 ) are also continuously three key figures for the neighbor ( 70 ) B who determines this transaction message ( 40 ) sent:

1. 1. Novelty value ^l transaction _n ',
2. 2. Validity 'Transaction _v ', and
3. 3. Change 'Transaction _c '

of transaction messages received from B ( 40 ).

verwendet.These values are used in the main work step "Network formation" (190) for the calculation of the rating ( 80 ) from the neighbor ( 70 ) B $\text{'}{R}_b^a\text{'}$

used.

In the subtask "Answer" of the main step "Finding Consensus" (170), the computer system replies ( 10th ) A to the neighbor ( 70 ) B the acceptance of the transaction message received ( 40 ) X. If A is the transaction message ( 40 ) Holds X as currently true, the content of the transaction message ( 40 ) approved. However, if A is another, the transaction message ( 40 ) X contradicting transaction message ( 310 ) Z holds true. Z is sent back to B instead. B can now decide whether to change his mind or not.

In any case, a response to a transaction message ( 40 ) from B no request from TbPoW (240) by BDh A does not have to solve a TbPoW (240) for this answer (see 6b ).

In any case, a computer system notices ( 10th ) which transaction messages ( 40 ) to which neighbors ( 70 ) has already sent. Infinite loops in the blockchain computer network ( 60 ) are also avoided by the fact that the same transaction message ( 40 ) at the same neighbor ( 70 ) is not sent more than once.

In the subtask “forwarding” of the main work step “finding consensus” (170), transaction messages received ( 40 ) to all other neighbors ( 70 ) forwarded. No transactional messages ( 40 ) which the computer system ( 10th ) A considers invalid. Also no transaction messages ( 40 ) forwarded to the respective neighbors with the same content ( 70 ) were sent.

For the sub-task "acceptance" (see 8th ) of the main work step "finding consensus" (170) a computer system is waiting ( 10th ) A that a) the news rating ( 90 ) a transaction message deemed valid by A ( 40 ) X reaches a minimum value, in particular because A receives this transaction message ( 40 ) also from other neighbors ( 70 ) with higher rating ( 80 ) receives, and that b) a certain time T1 passes without anyone else, the transaction message ( 40 ) X contradicting transaction messages ( 310 ) arrive. If both conditions are met, this transaction message ( 40 ) recognized as final and valid by A: A builds the change transaction (s) ( 30th ) from the transaction message ( 40 ) in your own version of the distributed blockchain database ( 20th ) a. The transaction message ( 40 ) will be contradicting you, if necessary Transaction messages ( 310 ) from the temporary local database ( 50 ) deleted.

If one in the temporary local database ( 50 ) transaction message accepted as currently valid ( 40 ) after the time has expired T2 not the desired minimum message rating ( 90 ) is reached, together with other transaction messages that contradict it ( 310 ) from the temporary local database ( 50 ) deleted. This is necessary as additional protection against SPAM ( 260 ) and for the completion of the temporary local database ( 50 ) to prevent.

The following always applies: 'T1 <T2'.

The described process of the main work step "finding consensus" (170) together with the results of the main work step "network formation" (190) result in the blockchain computer network ( 60 ) with contradicting transaction messages ( 310 ) Waves of negotiation communication that quickly settle down to a single valid “opinion” - this has reached consensus and the blockchain database ( 20th ) will be brought up to date in each case.

Reference symbol list

• 10th

Computer system: a computer device capable of connecting to other computer systems via the Internet or other networks

• 20

Blockchain database: a distributed database that is available as a complete or incomplete copy on multiple computer systems ( 10th ) exists and from chronologically ordered change transactions ( 30th ) exists (based on a principle similar to a general ledger), which are linked together using cryptographic methods

• 30

Change transaction: a description of the change in the current database of the blockchain database ( 20th ); this transaction will, if the respective computer system ( 10th ) accepted, usually attached or “chained” to one end of the cryptographic chain (s) of transactions

• 40

Transaction message: a communicated message between the computer systems ( 10th ) that have one or more change transactions ( 30th ) may include

• 50

Temporary local database: a local database or memory in a computer system ( 10th ), separate from the blockchain database ( 20th ), where transaction messages currently being processed ( 40 ) can be cached for a short time

• 60

Blockchain computer network: a network of computer systems ( 10th ), which is a common decentralized blockchain database ( 20th ) and their change transactions ( 30th ) manage and exchange transaction messages ( 40 ) change

• 70

Neighbor: a computer system ( 10th ) B, a computer system ( 10th ) A is known and with which a network connection is established once or regularly

• 80

Rating / trust: a numerical rating (e.g. between 0.0 and 1.0) that a computer system ( 10th ) A by itself via another computer system ( 10th ) B in a blockchain computer network ( 20th ) has continuously calculated or defined

• 90

Message rating: The evaluation of the possible validity of a transaction message ( 40 ) by the ratings ( 80 ) all neighbors ( 70 ) is calculated, which this transaction message ( 40 ) with exactly the same content to the computer system ( 10th ) having sent

• 100

Incoming connection: a connection from the perspective of the computer system ( 10th ) A, between the computer system ( 10th ) A and the computer system ( 10th ) B of a blockchain computer network ( 60 ) by the computer system ( 10th ) B was initiated

• 110

Outgoing connection: a connection from the perspective of the computer system ( 10th ) A, between the computer system ( 10th ) A and the computer system ( 10th ) B of a blockchain computer network ( 60 ) by the computer system ( 10th ) A was initiated

• 115

Identification: the identity of two computer systems ( 10th ) is checked with the public keys of the respective key pair from the public key encryption process and a corresponding handshake

• 120

Slot: a virtual connection point for an incoming ( 100 ) or outgoing ( 110 ) Connection that a computer system ( 10th ) with which other computer systems ( 10th ) of the blockchain computer network ( 60 ) can connect to transactional messages ( 40 ) and other data packets with this computer system ( 10th ) exchange.

• 130

Group: any computer system ( 10th ) in a blockchain computer network ( 60 ) can define and communicate its membership in one or more groups, this can affect both preferred connections to other computer systems ( 10th ) as well as the exchange of relevant transaction messages ( 40 ) to have

• 140

Personality: a random value (e.g. between 1 and 7) in a computer system ( 10th ) A, which is determined for a certain time window and leads to an active data connection between the computer system ( 10th ) A and a neighbor ( 70 ) B only remains as long as both have the same random value for the personality; Also incoming connections are accepted or rejected according to this principle

• 150

Epoch: a window of time that a computer system ( 10th ) during which there is a certain personality ( 140 ) having; after the end of the epoch a new epoch begins with the determination of a new personality ( 140 )

• 160

Contact list: all or a subset of all computer systems ( 10th ) that each computer system ( 10th ), identified by their respective public keys and the last known IP address

• 170

Finding consensus: an autonomous evaluation and selection of transaction messages ( 40 ) through any computer system ( 10th ) in a blockchain computer network ( 60 ), which serves to decide which change transaction (s) ( 30th ) in the blockchain database ( 20th ) should be included; ie also an assessment of the validity of the change transaction (s) ( 30th )

• 180

Multi-agent system (MAS): a system consisting of several similar or differently specialized acting units, software agents or computer systems ( 10th ) that collectively solve a problem, in the present patent this problem is the finding of consensus ( 170 ) and ensuring the integrity of a distributed database; one also speaks of a distributed artificial intelligence

• 190

Network formation / autopeering: A process for automatic finding and for networking or for establishing a connection to a computer system ( 10th ) with other computer systems ( 10th ) in a blockchain computer network ( 60 )

• 200

Network topography: A virtual representation of the entirety or a subset of the blockchain computer network ( 60 ) where the existing computer systems ( 10th ) and their connections to one another.

• 210

Nonce: The result of solving a specific cryptographic puzzle that relates to a specific piece of information or data packet; this result can easily be checked using a hash function

• 220

Proof-of-Work (PoW): Solving a specific cryptographic puzzle by a computer system ( 10th ) or his search for a specific “Nonce” result ( 210 ), mostly using many hash operations on a certain piece of information or data packet; this search usually requires a relatively long and heavy CPU load

• 230

Nonce difficulty: The difficulty of the PoW puzzle ( 220 ) which is a computer system ( 10th ) for a specific piece of information or data packet, e.g. a transaction message ( 40 ), must solve

• 240

Trust-bound-proof-of-work (TbPoW): unique PoW ( 220 ) where the nonce difficulty level ( 230 ) depending on the rating ( 80 ) a computer system ( 10th ) is determined: the higher the rating ( 80 ) the smaller the required level of difficulty

• 250

Malicious computer system: a computer system ( 10th ) in the blockchain computer network ( 60 ) that either a) the rules of the blockchain database ( 20th ) or the computer network's own interests would like to break through one or more attack scenarios, or b) due to technical defects a faulty and for the blockchain computer network ( 60 ) has harmful behavior

• 260

SPAM: the deliberate or unintentional communication overload of a computer system ( 10th ) with transactional messages ( 40 ) which valid or invalid change transaction (s) ( 30th ) may include; it is often an attack scenario

• 270

Eclipse: an attack scenario in which malicious computer systems ( 250 ) try all connection slots ( 120 ) a computer system ( 10th ) and thus the sovereignty over its communication with the rest of the blockchain network ( 60 ) to get

• 280

Sybil: an attack scenario in a blockchain computer network ( 60 ) by undermining or falsifying trust or reputation by creating false identities

• 290

Double spending: an attack scenario and / or an impermissible state in asynchronous distributed data processing in the blockchain computer network ( 60 ), where conflicting issues in the blockchain database ( 20th ) can arise; e.g. with a blockchain database ( 20th ) With money accounts, double spending could mean that an account has a negative balance because an amount of money from the account was illegally spent twice.

• 300

Attack load: the sum of all attack scenarios occurring simultaneously or staggered, executed by one or more malicious computer systems ( 250 ) against one or more computer systems ( 10th ) in a blockchain computer network ( 60 )

• 310

Contradictory / contradicting transaction messages: multiple transaction messages ( 40 ), which at the same time affect the same facts (eg accounts) but cannot be valid at the same time because they contradict each other in the facts; these can arise from "double spending" attacks (290) or other competing circumstances in a distributed asynchronous data processing process.

• 320

Communication control / firewall: a process that uses TbPoW (240) to overload a computer system ( 10th ), especially through SPAM attacks ( 260 ), prevented

Claims (8)

Working method of computer systems (10) in a blockchain computer network (60) to protect the integrity of a distributed database, characterized in that a) the computer system (10) continuously calculates the rating (80) and other key figures for the known neighbors (70) , which are decisive for the further interactions with each of these neighbors (70), and b) all connections to the neighbors (70) are re-established and / or disconnected after a dynamically determined or a fixed period of time and the rating (80) of each Neighbors (70) and other key figures, but possibly also a random variable are decisive for the connection establishment or confirmation, and c) the neighbors (70) exchange parts of their contact lists (160) with each other, and the respective rating for the scope of the exchange ( 80) and other key figures are decisive, and d) the computer system (10) can inquire from its neighbors (70) the rating (80) about a specific neighbor (70) and this is used for the own calculation of the rating (80) for this neighbor (70), and e) in the case of incoming transaction messages (40) the computer system (10) provides a firewall (320) which the respective neighbor (70) provides a " Trust-bound-proof-of-work ”(240), which relates to the data content of the transaction message (40), the public keys of both computer systems (10), and possibly a random variable defined by the receiving computer system (10) , and whose nonce difficulty level (230) depends on the rating (80) of the neighbor (70), the network and the CPU utilization of the computer system (10), and f) the computer system (10) incoming and outgoing transaction messages ( 40) temporarily stored in a temporary local database (50) for a short time, as long as they have not been finally accepted as valid in the blockchain database (20) or rejected as invalid, and g) the computer system during consensus building (170) (10) evaluates the incoming transaction messages (40) with the ratings (80) of the respective neighbors (70), and, in the case of contradicting transaction messages (310), the respective message rating (90) is decisive for the selection of the valid transaction message (40) , and the sender confirms the transaction message (40) received, rejects it, or sends back the alternative contradicting transaction message (40) which the computer system (10) currently considers to be more valid, and h) incoming transaction messages (40), provided that this is from the computer system (10 ) are recognized as valid at the respective point in time, are forwarded to other neighbors (70) and a transaction message (40) with the same content can never be sent to the same neighbor (70) several times within a certain time window, and i) a transaction message (40) is recognized as valid and final by a computer system (10) and its change transaction (s) (30) are recorded in the blockchain database (20) of the computer system (10) as soon as the message rating (90) of this transaction message (40 ) reaches a predefined level and as soon as a certain time passes without the new contradicting transaction message (310) is obtained from the computer system (10). Working procedures according to Claim 1 , characterized in that j) the key figures calculated in method step a) are derived from the rating (80) for each other computer system (10), the available incoming slots (120) for new neighbors (70), the group membership (130), the Connection quality, and personality (140), which is a random value from a small set of real numbers, and k) the rating (80) calculated in method step a) is determined from other factors, such as the duration of the relationship between the computer system (10) and a neighbor (70), the duration of active communication with this neighbor (70), the rating (80) of other neighbors (70) via this neighbor (70), the novelty value of the transaction messages (40) received Validity of the transaction messages (40) and the frequency of changes in the transaction messages (40). Working procedures according to Claim 1 , characterized in that 1) the connection establishment from method step b) to a computer system (10) can only be established via slots (120) and these slots (120) are divided into slots for incoming (100) and slots for outgoing connections ( 110), and the number of slots (120) for the incoming connections (100) is greater than the number of slots (120) for outgoing connections (110), and m) when establishing a connection from method step b) for an incoming connection (100) another computer system (10) must identify itself by means of its public key, the respective key pair from a public key encryption method, and a handshake, and n) when establishing the connection from method step b) the computer system (10) receives an incoming Connection (100) with a neighbor (70) taking into account all or a subset of the factors: rating (80) of the neighbor (70), his group membership (130), the Number of available incoming slots (120) for new neighbors and a random variable (personality) (140) of the neighbor (70), incoming, and o) when establishing the connection from method step b) a computer system (10) only outgoing connection (110) with a neighbor (70) takes into account all or a subset of the factors: the neighbor's rating (80) (70), its group membership (130), connection quality and its random variable (personality) (140), and p) the establishment of the connection from method step b) the computer system (10) has the goal of maintaining at least 5-10 active connections to neighbors (70) and, as long as this goal has not been achieved, it lowers the requirements for incoming connections (100). Working procedures according to Claim 1 , characterized in that q) in method step c) a computer system (10) can send a part of its contact list (160) to this neighbor (70) on request from a neighbor (70) and the scope of this exchange is determined by a maximum number, which describes the maximum number of contacts this computer system (10) would like to disclose to this neighbor (70), and r) in method step c) the maximum number of disclosed contacts from a computer system (10) to a specific neighbor (70) taking into account all or one Subset of the factors: rating (80) of the neighbor (70), its group membership (130) and connection quality, is determined, and s) in method step c) the contacts already communicated by a computer system (10) to a specific neighbor (70) are noted and this scope can only be expanded by the maximum number of disclosed contacts. Working procedures after Claim 1 , characterized in that t) in method step d) a computer system (10) can request its neighbors (70) their ratings (80) via another specific computer system (10) and their answers / opinions with the respective rating (80), which the computer system (10) has over the respective neighbor (70) is weighted, and u) in method step d) the weighted “opinions” or ratings (80) of the neighbors (70) about a specific computer system (10) are used to that the inquiring computer system (10) has its own “opinion” or rating (80) the other computer system (10) can form, about which it has asked the opinion of its neighbors (70), and v) in process step d) the computer system (10) periodically and continuously on the opinions of all neighbors (70) to all known to him Computer systems (10) in the blockchain computer network (60) informed. Working procedures according to Claim 1 , characterized in that w) in method step e) the computer system (10) accepts incoming transaction messages (40) by a communication controller (320) only as long as this transaction message (70) a response to another from the computer system (10) to this neighbor ( 70) outgoing transaction message (40) or as long as the neighbor (70) has a correct answer to a special TbPoW request (240) made by the computer system (10) and relating to this new incoming transaction message (40), ie nonce (210), and x) in method step e) the TbPoW (240) required by the computer system (10) explicitly relates to the hash content of the transaction message (40) received from the neighbor (70), the two obtains public keys of the sending neighbor (70) and the receiving computer system (10), and a random number determined by the computer system (10), and y) in step e) that of the computer system tbPoW (240) required by a neighbor (70) has a variable nonce difficulty level (230) and this taking into account all or a subset of the factors: the neighbor's rating (80) (70) , Average network load expressed as the average number of transaction messages received (40) per neighbor (70) per time unit, average number of transaction messages (40) received from this neighbor (70) per time unit, current CPU load on the computer system (10), continuously determined and z) a computer system (10) received transaction messages (40) are only considered in finding consensus (170) (in method step g)) as soon as they have overcome the firewall or communication controller (320) in method step e), Working procedures according to Claim 1 , characterized in that aa) in method step g) a computer system (10) contains the transaction messages (40), their content, their current message rating (90) and the identifying data of the neighbors (70) which contain this transaction message (40) with this content sent, temporarily stored in a temporary local database (50) (method step f)) as long as this transaction message (40) was not finally accepted or rejected, or this transaction message (40) is received immediately after receipt due to factual or other defects rejected, and ab) a computer system (10) evaluates each transaction message (40) with the rating (80) of the sender neighbor (70) and thereby receives the message rating (90), unless this transaction message (40) with the same content already exists in the temporary local database (50) (stored in method step f)), as a result of which the ratings (80) of all neighbors (70) that this Tr sent transaction message (40), used to determine the message rating (90), and ac) in method step g) a computer system (10) with two contradictory but syntactically and factually correct transaction messages (310) always for the validity of the transaction message ( 40) decides which currently has the highest night rating (90), and ad) a neighbor (70) can change his or her validity of several contradictory transaction messages (310) and the computer system (10) thereby from this neighbor (70) in the procedural step g) can receive several contradicting transaction messages (310) one after the other, which means that the computer system (10) also evaluates these “changes of opinion” by subtracting the message rating (90) slightly when evaluating the respective subsequent transaction messages (40) , and ae) in method step g) a computer system (10) is switched on Transaction message (40) replies to the sending neighbor (70): either with rejection, due to factual or formal defects in the transaction message (40), or with the current consent by also confirming the content of the transaction message (40), or, if that Computer system (10) currently considers another contradicting transaction message (310) in the temporary local database (50) to be more valid, it is instead communicated to the neighbor (70) as an alternative, and af) in method step g) by the computer system (10) the receipt and evaluation of incoming transaction messages (40) from a neighbor (70), statistics continuously updated for this neighbor (70) (which will later play a role in the next run of method step a)), in particular the novelty value of the transaction messages (40), which indicates what percentage of all transaction messages (40) received from this neighbor (70) was new to the computer system (10) The validity of the transaction messages (40), which indicates what percentage of all transaction messages (40) received from this neighbor (70) were finally recognized as valid by the computer system (10) and the frequency of changes, which indicates what percentage of all of this neighbor (70) received transaction messages (40) this Neighbor (40) later changes his mind and instead chooses other conflicting transaction messages (310). Working procedures according to Claim 1 , characterized in that ag) received transaction messages (40), if these are currently recognized as valid by the computer system (10), are forwarded to all other neighbors (70) in method step h), and ah) the computer system in method step h) ( 10) the contradictory transaction messages (310) are linked to one another in the temporary local database (50) and changes in opinion about the current validity of the respective transaction message (40) are also registered, and ai) in process step h) the computer system (10) sends transaction messages (40) and the identifying data of the recipient neighbors (70) are cached in the temporary local database (50) and thus never sends the same transaction message (40) in succession with the same content to the same neighbor (70) within a certain time window, except for the computer system (10) had a change in opinion in the current validity of the respective transaction message (40) registered, and aj) the temporary local database (50) functions on a document-based basis according to the NoSQL principle and the hash values of the contents of the stored transaction messages (40) act as access keys, and ak) in process step i) that they are valid viewed transaction messages (40) in the temporary local database (50), which do not receive any new conflicting transaction messages (310) after a certain time T1, but do not achieve a desired degree of message rating (90), after a certain time T2 from the temporary local database (50), with possibly linked, contradicting, transaction messages (310), are removed, wherein T2 is always larger than T1.

Images