by Asif Qumer Gill
University of Technology Sydney, Australia.
Abstract
The connectivity or loop between the architecture-operations is referred to here as the ArcOps, which is important for enabling the mapping and flow of information between architecture and operations for effective decision making and value realisation. The ArcOps pipeline aims to provide much needed end-to-end architecture design, execution, governance, observability, provenance, and traceability. This article discusses the ArcOps pipeline and toolchain examples or options from the information-driven adaptive EA (AEA) framework. The ArcOps pipeline uses data, analytics, and intelligence ( AI, generated AI, machine learning) capabilities for sourcing architecture data from operations, then converting it into actionable information for decision making about transition and future states of the enterprise architecture including roadmaps, implementation, assurance, governance and investment plans. The AEA framework has been published in the “Adaptive Enterprise Architecture as Information: Architecting Intelligent Enterprises” book: https://doi.org/10.1142/12961.
Keywords: Adaptive Enterprise Architecture, ArcOps, Data, Analytics, Artificial Intelligence, Machine Learning, Decisions
1. Introduction
Enterprise architecture (EA) is a key enabler of planning, implementing and governing changes in the design of structure, behaviour, value proposition and purpose of an enterprise for adaptation. Being adaptive refers to the ability to handle changes for desired efficiency and innovation. Efficiency refers to sustaining and improving the enterprise. Innovation refers to continuously growing and transforming an enterprise. Traditional slow and heavy documentation driven EA practices are challenged by the fast pace digital adaptation and decision information needs. Further, in the current context of digitalisation, while the EA can be used for planning digital initiatives, it is also important to look at transforming the traditional EA practice itself into an information-driven adaptive EA (AEA) for architecting digitally enabled intelligent enterprises. The AEA framework has been published in the “Adaptive Enterprise Architecture as Information: Architecting Intelligent Enterprises” book (Gill 2022a).
My earlier article “Adaptive enterprise architecture practice – data to decisions” (Gill 2022b) discussed the adaptive enterprise architecture as information (AEAI) to support adaptation decisions and actions for desired efficiency and innovation (outcomes and impacts). The information driven AEAI approach uses data, analytics, and intelligence (e.g. Artificial Intelligence, Generated Artificial Intelligence, Machine Learning) for architecting intelligent enterprises. This article focuses on the connectivity of enterprise architecture (Arc) and operations (Ops), which are often considered disjointed capabilities of an enterprise. Architects spend significant time in modelling the perceived or assumed current state of the architecture information via strategic stakeholders’ workshops while having limited to no use of the information from the Business and IT operations. Operations present the true and dynamic current state of the enterprise architecture in use, which must be looped back to the enterprise architecture and vice versa. The connectivity or loop between the architecture-operations is referred to here as the ArcOps, which is important for enabling the mapping and flow of information between architecture and operations for effective decision making and value realisation. This article discusses the ArcOps pipeline and toolchain examples or options from the information-driven adaptive EA (AEA) framework.
2. The AEA Framework
The AEA framework is organised into six interacting architecture information layers or domains: interaction, human, technology, facility, environment, and security. The AEA focuses on the information-driven adaptive capability to continuously scan and sense the internal and external environment architecture across layers for data about known and unknown events (complex event processing), changes or disruptions, interpret and analyse the collected data, and decide and respond based on intuition and rationale to expected and unexpected changes (threats and or opportunities) for adaptations (Figure 1) (Gill 2022a). This refers to the use of data to decision in the ArcOps approach in the AEA framework, which involves sourcing architecture data from operations, then converting it into actionable information for decision making about transition and future states of the enterprise architecture including roadmaps, implementation, assurance, governance and investment plans.
Figure 1. The Adaptive EA Framework (Gill 2022a)
3. The ArcOps
The ArcOps is an integrated pipeline of enterprise architecture (Arc) and operations (Ops) that aims to provide end-to-end architecture design, execution, governance, observability, provenance, and traceability. ArcOps is not an alternative to the existing AIOps, BizOps, DevOps, DataOps, MLOps etc. practices, rather ArcOps is an information-driven overarching pipeline, which can encompass these existing practices as a sub-set of the ArcOps. The ArcOps pipeline can be supported via different tools or toolchain. ArcOps is organised into 3 key parts: Architecture, Operations and ArcOps information fabric or enterprise knowledge graph (see Figure 2).
Figure 2. The ArcOps: Connected Architecture-Operations Pipeline
Architecture
Architecture data or information can be organised into six interacting architecture information layers or domains: interaction, human, technology, facility, environment, and security. The architect information can be stored in different architecture repositories or architecture information systems. Architects can use different tools for handling the architecture information. For instance, they can use Archi or Miro for generating the conceptual architecture information models and storing them in the architecture information systems or repository or library. They can also turn these conceptual models into data and store them in the property graph or semantic knowledge graph tools such as Neo4j or Stardog. This data can be further processed via the analytics tools such as Tableau and Power BI. Further, architects can also process the data using the AI/GAI/ML tools such as ChatGPT and MonkeyLearn. Architects can also use specialist EA software tools such as Abacus, Jalapeno or LeanIX to manage the architecture data and information. EA specific tools have started providing and/or integrating with some of the data, analytics and AI/GAI/ML capabilities or tools including operations (business-IT information systems), to support the information-driven ArcOps pipeline.
Operations
Operations data or information can be stored in both business and IT operations information systems. Business operations refers to actual business transactions, history and related data or information that can be stored in business specific enterprise systems (e.g. Salesforce CRM, SAP ERP), business application data stores (e.g. SQL Server, Mongo DB), data warehouses, data lakes, data lakehouse (e.g. Databricks) and business intelligence reporting information systems. IT operations refer to IT service management, which refers to managing of hardware and software assets, services and associated operational support and maintenance functions (e.g. help desk, configuration management, incident management, problem management). Operations information can be stored in different IT operations information systems such as configuration management databases (CMDBs), asset registers and service catalogs etc. There are several IT operations tools such as ServiceNow, JIRA and Confluent etc. The connected business and IT operations provide the true current state of the architecture related assets and services in operations. Such information can be linked to architecture information via the enterprise knowledge graph.
The ArcOps: Enterprise Knowledge Graph
The ArcOps information fabric or enterprise knowledge graph is a semantic metadata layer (Information Catalog) that connects the information residing in architecture and operations information systems or repositories. This does not require establishing a large and centralised ArcOps data warehouse. Rather, the architecture and operations information stay in their relevant repositories or information systems. However, the ArcOps information is connected via the semantic metadata driven enterprise knowledge graph layer. There are several ways to map, integrate or connect the architecture and operations information elements. However, “Product Service” is a key information element that can serve as an integration point between the architecture and operations elements. The ArcOps pipeline can be further connected to AI/GAI assisted search and discovery tools to provides access to integrated ArcOps information. The data can be processed (interpret and analyse) using the data analytics and intelligence capabilities for generating the actionable information. The information can be used for making informed decisions (decide and respond) for adaptation. Like any other information asset, ArcOps information needs to be governed and quality assured.
4. Conclusion
This article provided an overview of the integrated ArcOps approach from the “Adaptive Enterprise Architecture as Information: Architecting Intelligent Enterprises” book. This provides new perspectives in terms of provisioning the operations information (current state architecture) for making adaptation decisions about the transition and future state architectures. It also pointed out the use of an enterprise knowledge graph layer to connect the architecture and operations information elements in the ArcOps. Finally, it is mentioned that the “Product Service” is a key information element that can serve as an integration or mapping point between the architecture and operations information elements.
References
- Gill, A.Q. (2022a) Adaptive Enterprise Architecture as Information: Architecting Intelligent Enterprises. World Scientific. https://doi.org/10.1142/12961.
- Gill, A.Q. (2022b). Adaptive Enterprise Architecture Practice: Data to Decisions. Enterprise Architecture Professional Journal (EAPJ).
About the author:
Asif Gill is a result-oriented academic cum practitioner with an extensive 20+ years’ experience in IT in various sectors including banking, consulting, education, finance, government, non-profit, software, and telecommunication. He is A/Professor & Leader of the DigiSAS Lab at the School of Computer Science, UTS. He is also director and founder of the start-up “Infoagility” (Adapt Inn Pty Ltd). His earlier professional experience in agile software development, solution architecture, information architecture, information security, and program management provided a strong foundation for later work in digital strategy, architecture, and solutions. He is recognised as a global leader and specialist in adaptive enterprise & information architecture for designing and implementing large scale digital data ecosystems and platforms. He received his Ph.D. and M.Sc. in computing sciences. He is also certified as a CISM, DVDM, DCAM, ITILv3, and TOGAF.
