Mining information systems and the importance of appropriate architectural designs as key enablers for MIS deployments

Transcription

1 Mining information systems and the importance of appropriate architectural designs as key enablers for MIS deployments by Freddie Huyser, Optron Abstract The key in implementing mining information (MIS) solutions, is the recognition that the reason for MIS solutions is the lack of architectural thinking in implementing source systems, the latter of which are usually chosen for their specific application, with no integration and therefore unable to use information for decision making. Keywords management information system, data landscape, system design, mining information system, architectural designs The case for management information systems (MIS) in South African mines Mining systems adoption Mining houses often obtain software from different vendors to attain the best-of-breed offering for each functional application area. For example, mine planning departments may purchase a production-planning package from one vendor and geoscience applications from another. Even enterprise resource planning (ERP) vendors offer numerous enterprise applications and claim that their integrated system is a greater solution; however all modules in an ERP system are rarely best-of-breed. The main reason for the selection is based on the accountability and responsibility that one has in your profession to deliver the best application to support your functional role in a mining value chain. Fig. 1: Key components of a management information system (MIS). What is not considered at the time is the impact of the solution on up or downstream processes that are dependent on the inputs or outputs of other processes, and for that matter system, data flows and formats. Some benefits that can be attained using MIS MIS reports help to take decision and action on certain items with fast turnaround times. MIS can help a mine or division of the mine gain a competitive advantage. Competitiveness is a function or process ability to do something better, faster, cheaper, or uniquely, when compared with rival mines in the market. Mining companies are able to identify their strengths and weaknesses due to the presence of revenue reports, 129

2 production and product performance, employees' performance record etc. Identifying these aspects can help a company improve its business processes and operations. It can give an overall picture of the mine company s performance. It can act as a communication and planning tool. The availability of production, product data and feedback can help the mine to align its business processes according to the needs of its market. The effective management of production data can help the mine to perform direct to market demands Why is an architectural design approach key? The purpose of a design is to describe your end state. The design reflects you goals, requirements and also constrains you could face. The process of a design attempts to balance the various forces affecting a MIS design. Considerations in a MIS design include: What source systems would be involved to obtain MIS data from? What and how does the data landscape look that will be used to obtain the data? What business processes are involved in collecting, generating, and executing the information needed for a MIS? What organisational design of roles and responsibilities are used to get the information, and what dependencies of responsibility and accountability are there? There are a numbers of building blocks needed for a design concept. The aim is to have a design that is not the optimal solution, but rather a design that best satisfies the trade-offs between the different considerations and building blocks. In architectural design, you attempt to make decisions to transform your design purpose, requirements, constraints and architectural concerns, which you could call architectural drivers as shown in Fig. 2. Fig. 2: Overview of the architecture design activity (Image Brett Lamb/Dreamstime.com). What is architectural design? Architectural design is the process of making decisions to achieve goals and satisfy requirements and constrains. The output of a design is a direct reflection of such a goal, which is delivered as a requirement and contains the 130

3 desired end state. System design, or for that matter MIS system design, can be compared to the design of a house where different styles of houses in different countries are adamant. Although a house design would be a design for a house, it is in fact the requirement of the house; in China it would be different to the one that would be required in Alaska due to the different conditions, requirements and constrains. MIS systems design involves making decisions, and working with available skills and material to satisfy requirements and constrains. Why is architectural design important? With any system design or implementation, the importance of having a clearly defined goal, or for that matter a design, is imperative if you think what the impact on project cost could be. Even in early project scoping, it will be necessary to have an architectural design that will enable key approaches for achieving architectural drivers that form the base of the work-breakdown structures, and the choices of technologies, skills and integration for the MIS. How does architectural design work? Before commencing with a MIS architecture design you need to ask yourself what you are doing and why. Although this sounds patently obvious because you have already identified the need for a MIS platform to unify information, the devil is in the details of the design to satisfy the requirements with all the constraints that will arise in the process. Architectural drivers need to be base-lined and managed throughout the development life cycle: Design purpose: Clarify the purpose of the design that you would like to achieve, by asking when and why it is needed. What would be the most important business goals of concern to the mine or business at the time? Although architectural design could be done for a project proposal to enable a project selection, determining project feasibility, schedule, and costing an initial architecture need to be created. Such an architectural design would not be very detailed, and its purpose is to understand and break down the architecture into sufficient detail that the units of work are understood, and hence may only be an estimate. Another approach of finding the applicable driver for the purpose of an architectural design is to create a detailed design as part of the development of the MIS. This is to satisfy requirements, guide system intention and construction and work assignments, and prepare for an eventual release of modular implementations of a MIS. The architect should be clear about these goals and should communicate them (not negotiate them) to establish a clear design purpose before beginning the design process. Quality attributes are defined as the testable and measurable properties of a system that could be used to test how well the system satisfies the needs of the stakeholders. Among the drivers of an architectural design, quality attributes are the ones that shape the architecture most significantly. The critical choices that are made when working on the architecture determine in a large part how your MIS system will or will not meet those driving quality attributes goals. The different design levels are in many cases over simplified by referring to a high-level design, but the real aspects of a design goes to the level of: Architectural design Elements of interaction design Elements of internal design Finally, a well-designed, properly communicated architecture is key to achieve agreements that will guide the project implementation team of a MIS. Some of the most important agreements are to be made on interphases for ETL layers, shared resources and infrastructure needed. Findings of recent mining system architectures in practice In a recent analysis of typical current mine systems architectures, it was highlighted that integration is the last, if at all, considered part in implementing the system landscape. Businesses become trapped in specialised business processes and are assigned with cross accountabilities for data, not withstanding the duplication and manual handling of data, just to facilitate a standard report or convert data format to start the next value chain process 131

4 processing. Methodology used for analysis Leveraging a complete analysis of a mine s core value chain activities will be fundamental to depict the current architecture of how the landscape of each dimensions is functioning at its current state: System architecture Data landscape Process architecture Organisation design and structures Furthermore, the importance of analysis on process connections, data flows, interactions and format to depict the business architecture landscape that evolved over time, will be required to highlight interoperability elements that show trends and relationships to the overall integrated mining value chain. Fig. 3: Data repositories and operational systems to support mining processes. Understanding the needs of a MIS system is clear when you have multiple data sources that are required over a large number of expert systems, which use different formats and file base methods to generate outputs that have not been designed/considered for integration which other systems. The overall architectural landscape could be presented as in Fig. 2, stating the need to connect data repositories that are gathered by expert systems on the MES layer through a high level of different and sometime duplicated manual reporting processes. Benefits to be unlocked from having a MIS The need for an integrated, unified data environment can be seen by acknowledging current shortcomings. Current state: Recapture/import the same information into different systems because systems don t communicate with each other Different people doing the same thing on different systems duplication of effort Many versions of the truth and low level of transparency Redo things because of faulty ( untrustworthy ) information received Spend a lot of time to manually change the format of information for reporting and transporting purposes Generate a lot of good data and information that we don t use inability to readily and consistently extract, 132

5 collate and integrate operational data and information across the value chain Wait for someone else to provide information before being able to continue with own work (information focus in-function instead of a cross-function) Future state: Improve the way we generate, collect, integrate, analyse, compare, identify trends and report data and information Reduce the need for repetitive manual data entry and conversion Have access to near real-time automated information with a first level of interpretation to see what is going on as it happens right information in the right format at the right time Have an integrated dashboard ( one version of the truth ) across the mine s value chain to improve our decision-making ability Plan better and react sooner and more effectively to changes in production conditions/outputs across the mine Improved stakeholder engagement due to better visibility and transparency, and an overall increase in turnaround time, quality and integrity of information Help you do things faster (more efficiently) and right the first time (more effectively) Reasons for architectural design before implementation of a MIS High level The following high-level observation was made during an analysis of all functional elements on a mine. This was done to establish a design architecture, to illustrate that the business requirements, the condition of the current data landscape, the way the business is executed and the assignment of roles and responsibilities were understood, and to enable the definition of an informed scope of what would be possible. The following typical points in the core mining value chain have highlighted the need for a MIS deployment: Timing of information was not always aligned to the business cycle requirements, e.g. during the parallel interdependent processes of key business validation phases, support processes and production could not get the required information in time to make decisions. The source system databases ensure good quality data with automatic checks and balances however flat file systems allow the users to still use the wrong version of the data. Employee capacity (numbers and expertise) add a degree of risk to long-term company data and information integrity. Timing of cycles between different planning processes and their reconciliation contribute to lag in availability of information to take corrective actions. Large number of incompatible systems necessitate data interchange and manual processing of data. Difficulty to maintain data integrity and even more difficulty to ensure effective version control resulted in large numbers of different data repositories or data storage versions. Significant amount of consultation is required across functions to generate management reports due to process, system and data owners being allocated different roles. Different accountabilities and responsibilities potentially compromise the ability to generate level specific information for management decision. The impact of not having a consistent, shared view and understanding of the business process, is the inability to clearly identify the cross-functional dependencies on information flow and to optimise the timing of the flow of information. Although most geospatial data processing was done in a common system, which limits data conversion, there were still the necessity to have a file-based proprietary system, which does not allow multi-user access. Additional responsibilities were added to existing roles to generate level-specific information for management decision making. 133

6 Some duplication of data and information is created through the core mining value chain processes. Availability of information to support the end-to-end business process is hampered due to timing of crossfunctional interdependencies. Incomplete data due to the use of spreadsheets instead of databases were visible across the core mining value chain. Data is not leveraged in full; some critical systems have meaningful information but it is not fully utilised. Although process and information flows have been used in an optimised way, there is some duplication of effort to manually maintain the Excel reports. With a focus on the fundamental drivers sections of a MIS, the following summary of key points illustrate the need for a MIS platform. Business process architecture The importance of the business process landscape pose potential risks from a process architecture perspective where the current processes were observed to be a combination of process maturity, handoffs between roles/functions, and handoffs between systems/technologies. Data architecture The utilisation of SQL with a reporting tool enables multi-user access to the latest data for efficient reporting. The manual and file base data conversions result in negative data flows that were hampered by file-based products/systems, which allow multiple versions of the truth to be created. Significant use of Excel in quality control processes increases risk for errors and manual manipulation. Some of the software do not use database technology, which result in a lot of importing, exporting and translating of data from one system to the next. Systems architecture The system landscape architectural viewpoint was that they were working with incomplete data due to the use of spreadsheets instead of databases. Automated transfer of data was not always possible due to some systems creating gaps and errors. The lack of integrated systems hampers data flows and interface between key production systems, while data not leveraged in full result in meaningful information not utilised. Data landscape Good utilisation of SQL with a reporting tool enables multi-user access to the latest data for efficient reporting. File-based products, e.g. Microstation and ArcGIS allows for multiple versions of the truth to be created. Significant use of Excel in e.g. grade control and plant increases risk for errors and manual manipulation. Some of the software do not use database technology, which result in a lot of importing, exporting and translating of data from one system to the next. Reports The lack of common reporting naming conventions, with reporting names changing month-to-month, impacts on ability of third parties to consistently interpret/scrutinise information. Searching for the most up to date reports is time consuming and cumbersome. There is no consolidated summary report spanning across the value chain, which inevitably delays decisionmaking ability. Loss of data integrity due to a number of duplicated reports and different reports containing the same data. 134

7 Roles and responsibilities It was made apparent that: Additional responsibilities were added to existing roles to be able to generate level specific information for management decision making; this also leads to duplication of data and information across functions. Data and information generated, integrated, and in some instances approved by one individual or small groups of individuals put data and information integrity at risk ( entertainer and arbitrator ). Ability to retain knowledge and IP limited lose organisational memory if individuals leave. There is no doubt that for the MIS to reach its full potential, which is expected to make a success of the mine, fundamental alternatives will have to be considered based on: A re-assessment of the critical asset value of data Governance structures and mechanisms Re-defining of roles and capability building Re-looking the optimisation of current technologies and data and technology architectures Re-defining and re-prioritisation of must have, accurate, automated and reliable MIS reports and MIS practices The solution leans towards a business transformation of MIS practice on a fundamental level rather than a couple of quick fixes of symptoms. This will ask for and be dependent on definitive shifts in the way mining leadership thinks about MIS and how the future changes will be owned and led. References Fig. 4: Example of business architectures depicting the complexity for an MIS. [1] Humberto Cervantes, Rick Kazman: Designing Software Architectures: A Practical Approach Chapter 9. [2] Juan Pablo: Big Data Analytics: Information 2020: Beyond Big Data. [3] Management information systems, Contact Freddie Huyser, Optron, Tel , fhuyser@optron.com 135