Epilogue

A Cognitive Distributed AI Virtual Machine for the Processing and Understanding of Data

Epilogue is an experiment in the creation of a distributed data and analytics appliance specifically for processing large geo-distributed data sets that are a hybrid of denormalized relational, network, graph and document (unstructured).

The system has at its basis the following technologies (in a somewhat tailored form)

·       A stripped-down embedded Linux kernel

·       A highly customized Erlang/OTP implementation (a domain specific alteration of BEAM)

·       An extended Prolog implementation (tightly integrated with BEAM and the Linux Kernel

·       A minimized embedded TCL implementation (tightly integrated with BEAM)

On top of this infrastructure analytics capabilities are built that span a large virtual memory and storage space that enforce data rules around (GDPR, and other governance and restriction requirements).

Epilogue should not be viewed as a data store but as a vault implementation of an object store where objects have intelligence relative to how they live within the environment (e.g. Objects are responsible for access, location, promotion, integrity, and evolution).

 

Immediate Work:

·       Mapping out the virtual architecture

·       Breaking out Linux services and mapping into BEAM (Erlang/Beam being the sole access to OS services).

·       Refactoring BEAM to work in this specific architecture with little to no other environment consideration where there would be any performance trade off.

See the documents directory for more information as the product progresses.