Mitigating security concerns

Private at heart and secure by design

As a user of the system your memories might live forever over the internet. The security, privacy and legacy of your own data and of your Ai Self stays at the heart of our vision and design and is a foundation element of our ecosystem development.#

To address certain flaws in the security architecture of the system the founding team decided to split the evolutoin of the technology in 3 phases, decribed earlier in the paper.

The foundation of the system is always connected to an order of chronology of things protected in 3 different user ledgers ( pre-cortex, block and memory meta-chain) and validated over and over again during the anti proof of history. -> More on Flow of a Raw file to a memory page.

FAQs

How do you avoid centralized storage privacy concerns?

All the files are stored as encrypted on the storage and they can be unlocked only with the unique wallet key of each user. As a continuity of this , we are developing in phase 2 a fully decentralized file split system that will replace the centralized storage.

How does the system avoid the wallet access being compromised due to improper authentication or social engineering?

As any system as long as it has full user permissions there are not many options to defend the integrity of the memories against. Although due to the fact that we included in our system design a synching mechanism between multiple wallets of the same user and because the user ledger contains a proof of history and anti proof of history verification mechanism the system has a quite high level of security and historical integrity.

Also the authentication from a secondary wallet has to be signed in the master wallet. If the master wallet is compromised the intruder will have full access to read, remove or add new memories, but thanks to the memory meta chain ledger time stamping of entries and due to the block ledgers time stamp of changes it can easily be seen what actions were performed and can be reverted.

How do you avoid alternation of historical memories via insertion of fraudulent files?

Due to the innovatie concept of factual and cognitive blocks, that contain a validation of the creation through the timestamp of the file, checked during the pre-cortex processing and validated from the pre-cortex ledger, the system knows if a file was processed or not in the past and what factual and cognitive blocks it contained. The pre-cortex ledger holds itself a historical chronology of when a file was added and what file and a return in time to a previous setup is easy to be made.

Also a file that is newly created, will have a recent time stamp, can be added to the memory, but will be processed accordingly in relation to the cognitive and factual blocks extracted from it and will appear as an enclave of uknown factual information in the chronological logged memories.

If that is edited visually and informationaly as a file that contains information relevant to the user personal models, respectively factual blocks and child hood learning, the system will be tricked to created relations between this and a new memory, but the same mechanism of the pre-cortex ledger will easily help to identify when a file was added, from what wallet and what blocks were extracted from it and if that is detected as fraudulent or ranked by the user as fraudulent, the anti-proof of history will re organize everthing excluding the flawn component.

Please read the lifecycle of a memory to understand the details of the blocks involved.

File encryption

Each memory has its own encryption key, network identifie and a split of files that compose it .

An encoded variant of the files is stored on the MegaNode for private computing Ai processing purposes.

Ai confidential node computing

The Ai processing nodes do not recieve from the wallet or from the storage reconstructable files but custom encoded packages uniquley understood only by the processing algorithm. This means no private data is exposed to the node during the factual and cognitive extraction and the computation happens within an encoded environment where encoding model is unknown to the public and can not be deconstructed due to the network controlled proof of usage.

The network keeps in a user ledger the proof of reconstruction of each memory and by whom and what point was used, as well as what was the purpose of the use.

Scalling storage with privacy and security at heart

ETHERNOS is a network of demand and offer in terms of storage and processing capabilities. It is foreseen that there might be storage scalablity network limitations due to missing enough community nodes to keep the storage in escrow. Due to this reason the Network has a buffer storage called MegaNode that acts as a in between storage space with huge capabilities that is run by Ethernos and not by the community.

The MegaNode is responsible for storing user memories for a limited period of time until enough decentralized nodes are available.

The MegaNode also acts as a temporary buffer for storing clones of encrypted pieces of memory files, which when stored on nodes that get offline for a certain period of time, get automatically replicated by the network in order to maintain the content persistancy and integrity.