The Orchestra platform is a suite of technology, tools, and code libraries purpose built to enable development teams to build "Knowledge Orchestration" systems. Knowledge orchestration systems are different from information systems in that at their core they focus on developing understanding and exerting control and by their nature are dynamic. Information systems, business intelligence and data lakes are about data aggregation and static analysis and focus more on observation.
The Orchestra platform originated in the late 1990s in the United States Intelligence Community and its Department of Energy Laboratories. Issues with organizing diverse and dynamic information sets at scale and enabling advanced analytic techniques exceeded the ability of traditional relational databases and query response systems and led to the formulation of large in memory graph analytic solutions. Later while at Georgetown University, Dr. J. Smart evolved these early concepts to enable, via the AvesTerra program, the ability to organize and analyze knowledge at a global scale. It aims to enable global-scale analysis without compromising individual privacy or necessitating centralized control. The decades of work which led to the creation of the AvesTerra are AVT technology - the basic hypergraph transaction protocol, an assoicated API and the peer to peer HGTP server is at the core of the Orchestra platform.
The foundational principles are based on the understanding the differences between data, information, and knowledge. The ubiquity of information and communications technology have created a situation where the volume and variety of data and information created are increasing at an alarming rate. Data, characterized as the raw 1's and 0's that move across cables, transmitted through the air, stored on disks, etc., or information, data placed in context as individual transactions, reports, etc, are simply out pacing our ability to store, move and make use of these volumes effectively. The key lies in abstraction. Modeling knowledge from data and information. Bulding knowledge systems not infomration systems. THis involve representation of real-world entities - not conforming massive volumes of transactional data and information to some gold standard scheme so it can be joined together and analyzed. If you think about it,the number of "entities," people, places, things, events, etc. grow asymptotically to some upper bound, while the volume of data describing these entities exhibits exponential growth.
This distinction is important because analysis tasks are often focused on a relatively limited set of real-world entity types and there exists only a finite number of instances of these object types. This contrasts sharply with the endless nature of data that describes various aspects of these entities.
Orchestra is the peer to peer network to create, access, and manage a shared global-scale knowledge space, which can be implemented on various computational infrastructures. By isolating the design and implementation of analytic applications from the underlying complexity of this diverse and ever-changing infrastructure, collaborative communities can develop analytic components at a global level without being hindered by low-level implementation details. Knowledge orchestration systems with Orchestra under the hood use the HyperGraph Transaction Protocol (HGTP) to communicate. The exact details of the underlying API used to communicate with knowledge space can be found in the AvesTerra API technical doc. Understanding of HGTP or API is not needed to understand how it works, Orchestra provides a friendly interface and does the hard work of managing the servers for you.
Simplistically a knowledge space is a set of entities distributed across a set of Orchestra "hosts". A host is another name for an HGTP Server. Each entity has a unique entity identifier which encapsulates its "address" - i.e. what host it lives on. Relationshps among entities are created when one fascet of the entity's model refers to another entity by its entity id. When an Orchestra server cannot directly fulfill an API request due to the referenced entity being managed elsewhere, it forwards the request to the appropriate remote server for processing. The results are then forwarded back to the server so that the original request can be fulfilled.
The knowledge space is a collection of entities that represent instances of various concepts and relationships. These entities possess attributes that characterize, relate, and differentiate them from each other, many of which may change dynamically over time. Additionally, entities have "agency" in Orchestra and can come with methods enabling various actions to be performed upon them. All users of Orchestra have access to the knowledge space, subject to authorization restrictions determined by the knowledge originators.
The primary interface constructs for interacting with this knowledge space include:
Orchestra is ontologically neutral, however in order to convey semantic meaning, we need to agree on a Taxonomy to use. The taxonomy is the language spoken by the system. Everyone using English agrees what the meaning of the word 'horse' is. In Orchestra, everyone using the same taxonomy agrees what the meaning of the Attribute 423 is. This taxonomy includes, but is not limited to, the Class and Subclass present in the entities metadata, the list of Method adapters may have, the list of Attribute used to describe things etc...
Think of Orchestra as a layer on top of the core entities that brings a standard taxonomy, as well as some standard adapters to handle common operations. It also defines how data is structured. Orchestra defines what types of values there are, and how they are encoded in bytes.
Orchestra also defines how data is stored in entities, with a data type called the Orchestra Model.
Orchestra also provides additional utilities required to build and operate knowledge orchestraiton systems:
Finally avu is a Command Line Interface (CLI) tool included in every Appliance that allows one to interract with a knowledge space by viewing entity data, see their metadata, manage authorizations, create new entities, change entities authorizations and much more.