Dereks Notes.

From The Central Electricity Generating Board Diary for 1987. Notes Written on pages for October 15th to October 21st. Derek's Notes were always written in a cramped, but legible, handwriting. He preferred the security of a sturdy notebook to digital storage. Which was always a problem. His entire collection of notebooks could kill a man - a large man - if they fell on him.

The Distributed Ledger is a consensus of replicated, shared, systematized data distributed across a network. As a common implementation of distributed ledger technology, a Blockchain is one of the most widely-known types. Time is my network. Each node is an oasis of fluctuating entropy. As entropy rises, time moves towards that space-time location. As entropy falls, time moves away from that space-time location. The secret of time travel is to pick moments in space-time that are easy to generate entropy.

I had always thought that Derek was writing fiction. It turns out that his ideas on Time Travel were actually quite practical. They all centred on moving between puddles of different levels of entropy.

The main purpose of a Distributed Ledger is to serve as a digital record. That is it. Nothing fancy. Just a record of digital things. The power of the Distributed Ledger comes from that Record. For cryptocurrencies, a distributed ledger Records ownership by storing the details of all Property Exchange Transactions publicly across the network. The reality for Time Travellers is that the exchange of ownership can only take place at the instant that their timelines cross.

And his notions about the requirements and obligations that fall to the average time traveller would not be out of place in a physics text book.

The Ledger, like Time, require no central administrator, but that is a sufficiency of omission not a necessity of prohibition. The Ledger is replicated among the nodes in a peer-to-peer (P2P) network. For Cryptocurrencies the Bitcoin Network is the definition of Bitcoin itself. To gain perspective on what this means, placing the Ledger into the context of the International Standards Organisation (OSI) Seven Layer Network Model:

Layer 1: Physical, (bit, symbol).
Layer 2: Data Link, (frame).
Layer 3: Network, (packet).
Layer 4: Transport, (Segment, datagram).
Layer 5: Session, (Data).
Layer 6: Presentation, (Data).
Layer 7: Application, (Data).

The Seven Layer Model appears, to Engineers to be all about hardware intercommunication. In Layer 7, Engineers acknowledge that there is some kind of "end user role" and suppose that lower level Layers are of no interest to the "end user". This is like Time Travel, everybody wants to be able to push a button and be some place, some time without ever worrying about the detail of how the process works. Layer 3 concerns itself with transferring packets from hardware node to hardware node. Layer 3 does not guarantee reliable delivery or have a necessity for reporting message delivery or message delivery failures. People forget that the ISO Model is not about perfection but expediency. Time Travel needs perfection and expedience. It is a whole different prospect to building computer networks. One of the significantly coincidental areas is trust.

Derek was never really given to explaining why he chose the analogies that he did. But when he chose an analogy it was always for a good reason.

The unreliable delivery being invisible to the Application, at Level 7, could potentially be a problem for Layer 7 trusting Layer 3. Which highlights that the Ledger is not simply a digital data pattern of the Ledger. The underlying Layers - beneath the Application Layer 7 where the Blockchain would be supposed to reside in a purely Engineering viewpoint - are an important element of whatever the Distributed Ledger purports to record. The Network is not simply a passive medium.

Peppering his notes with random quotes from people as varied as Marshall McLuhan and Max Stirner, actually extracting the ideas Derek jotted down can be a trial:

The medium is also the message.

So I, inevitably, stick to the technical pieces that I can understand.

The Application Layer - Layer 7 - implements a "New Network Layer". This New Network Layer is recursively reliant on all the layers beneath it. Without debating the Engineering details of each layer, that means that adding layers does not increase anything except the number of layers that the "New Network Layer" is dependent upon. Which leads to a serious conceptual problem.

The Bell-La-Padula Model (BLP) is a State Machine Model used for enforcing access control in secure Applications. It is not a particular technology but a set of principles.

The Simple Security Principle: a subject at a given security level may not read an object at a higher security level.
The Star Security Principle: a subject at a given security level may not write to any object at a lower security level.
The Discretionary Security Principle: uses an access matrix to specify the discretionary access control.
The Tranquility Principle: the classification of a subject does not change while being referenced.

The general summary for the BLP Model is: "Write Up. Read Down." (WURD). The Star Principle can be strengthened to only permit reading at the same level. Normally, the entire scope of the BLP is limited to Application Layer 7. However there is a more revealing way to examine it.

Which leaves my head spinning. Derek would often be off, ahead, talking seven ideas ahead of the idea you were assimilating. The kind of flurry of insight that looks just like madness.

Considering this in terms of the OSI Seven Layer Model: a Ledger at OSI Level 7 is a subject at the lowest level of the BLP Hierarchy. This is not difficult to explain: as you go from Layer 1 upwards, each layer becomes more complex, with an increased number of interacting elements. That increase in the number of elements increases the number of interactions and so the number of potential errors. There is no need to analyse those errors in terms of malice: the number of errors would increase in an utterly benign environment. As you rise through the hierarchy, the error increase is also an entropy increase.

When considering if a "New Network Layer" - the Appplication Layer 7 network, built in software, that constitutes the Distributed Ledger - is secure, there is a consideration that the Distributed Ledger directly draws from Layer 5, Layer 6, and Layer 7 to provide the Distributed Ledger "New Network Layer". That means being able to draw on the BLP Principles to discuss the Trust that is invested into the Ledger. The "New Network Layer" is always in a state of entropic flux.

Within Distributed Ledgers, the Consensus Algorithm seeks to ensures that each node copy of the ledger is identical to every other node copy. Which is, essentially, an entropy balancing process. This set of copies is called the Single Shared Ledger. Asset Owners use a Trust method: in effect the Tranquility Principle in operation. This Trust method is usually a cryptographic signature at the point of transaction commitment. That transaction is then recorded on the Distributed Ledger and shared to all participating nodes of the network. The Distributed Ledger is owned, updated and monitored by everyone and controlled by no one. Time, eventually, becomes universally more entropic.

Data sharing on the Distributed Ledger is shared and synchronised consensually among the nodes. The participant nodes of the network are required to agree about the state of the ledger periodically in order to protect the system and to keep it in check. Such an agreement is reached automatically thanks to a Peer-to-Peer Network intended to be transparent. Network monitoring for errors - either malicious or not - is possible by any node. Only when the Peers agree does the Distributed Ledger gain Consensus.

Remember Layer 3: there could be errors at that level - say flipped bits - that could be accepted, or rejected, at Level 7. This would be a matter of Consensus. Where the Digital Assets on the Distributed Ledger are simply representative of fungibility, this is an easy matter to be resolved. There are two states: fungible or not fungible. The detailed analysis of the not fungible subjects are not necessary because Consensus simply needs to discriminate between fungible (accept) and non fungible (rejects) and, there is a generally accepted idea of what fungible is and why fungible is desirable and so little analysis goes into the fungible. This results in the rejection of that which is not accepted as fungible - which eliminates any need to know of the vagueness of non fungible items.

Which, if you analyse from the perspective of a Time Traveller becomes interesting and worrying.

Generally speaking, Distributed Ledgers are regarded as either permissioned and unpermissioned. Unpermissioned Ledgers allow anyone to participate without being owned by a single authority, whereas the permissioned ledgers have a centralised, owning, authority and mechanisms for including and excluding participants. The reality is that it should be possible to have a centralised but unowned authority: which is the role that Time can take. Over time, even unpermissioned Ledgers become permissioned by things such as "market sentiment". The noise of the market being akin to those hypothetical Level 3 bit flips that could compromise the imputed integrity of the Seven Layer Model from a BLP perspective.

When the lessons of the OSI 7 Layer Model are applied to Cryptocurrencies over Time, there are possibilities. There are hints that there is something wrong with Time.