Enterprise Data (data@scale) can be said to be broken down into four dimensions; Volume, scale of data; Velocity, of streaming data; Variety, different forms of data; Veracity, uncertainty of data (IBM 2017).
Traditional techniques are struggling to deal with the scale, incidence, variety and quality of data@scale which has led to disillusionment in many projects (Gartner 2016).
Key factors identified as inhibitors are ‘data quality and reliability related issues’ and ‘not establishing corporate ontology and definitions for single version of truth culture’ (Analytics 2016) and failing to make sense of data other than rdbms.
Exzession Matrix is a Data Appliance which has at its core a Universal Ontological Model (UOM) addressing these challenges by creating a model driven “operational ontology” with a scalable technical solution for rapid deployment of analytics to expose “what you didn’t know you knew” and “allowing you to record things you didn’t know you needed to know”.
Exzession has developed disruptive technology used to deliver solutions that can massively reduce the amount of software, time and cost required to deliver analytics. It facilitates rapid change and moves corporate data from its current 2-dimensional state (relational databases) to being 4-dimensional information (space and time). This 4-dimensional nature provides unparalleled query and analysis capability for data@scale. Exzession believes this is the information technology missing link, “Business Information Fabric”.
It can be used to create the following systems; greenfield; new information from legacy data; and adding new life to legacy data; with an informational adjunct from the existing data. It can sit alongside bespoke or commercial systems to add the “information piece” that cannot be engineered-in or queried-out.
It enhances the data@scale proposition by adding meaning and data-type classification in support of and to enlarge the work by data scientists. This is a unique way to look at the value of data and to uplift the utility of systems wishing to use the data. Information becomes ubiquitous and represented in a common and consistent manner irrespective of its type and origin.
Based on many of years of research, the UOM is set theoretically true (mathematically true and certain), stratified and refactored for enterprise scale usage. A fundamental aspect of the UOM is that it is 4-dimensional, recording things, their relationship to other things, types of things and relationships (including types of relationships) and the time aspects of everything.
Client Business Models (CBM) plumb into the UOM providing a single view of all the CBM’s and “Semantic Interoperability” between them. Since CBM is directly descended from the UOM it is also set theoretically true. Enterprises will have many CBM domains which are likely to be derived from legacy systems and a feature of the Exzession implementations are the ability to expand the models in each domain (and across domains) over time.
Matrix implements Model Driven Information Architecture by reading the UOM and encoding it. The CBM is then also loaded and plumbed into the UOM.
Matrix then has a set of reusable encoded Micro and Macro orchestration patterns (derived from the UOM) and by reading the CBM identifies those patterns required to deliver an operational ontology. There are two types of patterns;
Micro Patterns – which are universal and business agnostic. They include; Recognition – Representation, Attribution and Identification; Set Inclusion; Set Membership; Partial Order – Parthood and Counterpart; Temporality – Intermittent, Subsequent and Phased; Events – Multiple Event Types and States.
Macro Patterns – which are business aware. They include; Temporality in Context, Geo Political Region, Located At, Asset Exchange, Unit of Measure, Multiple Classification, Equations.
Matrix then has a set of reusable encoded Micro and Macro orchestration patterns (derived from the UOM) and by reading the CBM identifies those patterns required to deliver an operational ontology.
Matrix/Patterns are encoded once and then Matrix/Orchestration combines them at runtime. This unique methodology results in the code base being extremely small and efficient. Real time pattern combination and actuation supports multiple and evolving CBM’s with no code changes.
Organisations are struggling to deliver value from their data@scale implementations. Inhibitors are ‘data quality and reliability’ and ‘not establishing corporate ontology and definitions for single version of truth’ (Analytics 2016).
Matrix/Factum is a configuration of the Matrix that is non-intrusive, has a Universal Ontological Model (UOM) at its core, can integrate data taxonomies from multiple models, provides mechanism for unified naming conventions, records lifecycle (data lineage) and can be deployed quickly by focusing on what some organisations call Critical Data Elements (CDE) whilst scoping future iterations to encompass wider data sets.
Cloud Storage products like Google and Amazon S3 store any file with a single key and an ability to add textual tags to that file. The single key is a meaningless reference, lose the key and you lose the file, the textual tags have no structure around them and suffer from Homonym’s and Synonym’s which will therefore give invalid results when analytics are run across data@scale environments.
Matrix/Index is a configuration of the Matrix/Factum providing a non-invasive approach to deliver an “index of anything and everything” application. Supported by the UOM, it provides semantically clean information to the analytic process considerably increasing the accuracy of the results and exposing “what you didn’t know you knew”. Queries can be run to retrieve files based on any quantum and/or their parents or children (transitive closure).
Regulators are stating that it is critical for enterprises to properly record changes to their data over its lifecycle. For example, the BCBS239 banking regulation states that “controls throughout the lifecycle of the data” are mandatory. Enterprise processes have become so complex over time that this is extremely difficult to achieve.
Matrix/Data Lineage is a configuration of the Matrix/Factum, supported by the UOM, which delivers an application that integrates current and legacy system models into a corporate ontology and provides the lifecycle of each object and the reasons for changes over their lifecycle, thus meeting a key regulatory requirement for many organisations.
* These fields are required.