Innovate decision making by combining Enterprise Architecture with Simulation Modelling

EA Dynamics UK is a training and knowledge transfer organisation keen to empower employees and increase organisational performance. This article reports on business uses of TOGAF and simulation modelling within global enterprises. 

Simulation modelling information  assists effective communication between business, IT and operations. Improving communications and understanding between business, operations and IT is high on the list for organisations.  There is a correlation between simulation modelling and improved decision making capability.

It is important to measure, manage and report service performance, periodically assessing alignment of IT with business goals as technology and business environments change. Modelling information provides valuable input into executive dash-boards.

 Known simulation uses include:
• Evaluating and looking for existing bottlenecks in processes.
• Collating historical and real-time data on service performance.
• Present the performance data necessary to make informed decisions concerning IT's ability to meet SLA's.
• Experimentation with multiple scenarios to determine resources needed to meet business unit goals.
• Establishing the source and root cause of outages.
• Averting availability problems from happening again.
• Proactively detecting potential availability issues before they impact users.
• Regularly tracking and availability metric reports .
 

BENEFITS

The benefits from effective modelling and simulation help the business to increase revenue through maximising availability, decreasing downtime, improving response times, increasing productivity, increasing responsiveness to market dynamics, increasing return on existing IT investment Decreasing costs through higher capacity utilisation, improving processes, just-in-time upgrades, and increased cost control.

The simulations and modelling supports IT and business alignment, helping to show the cost and business need for infrastructure upgrades.  Simulation and modelling can also be applied during the Architecture Development Method phases to support optimisation of the proposed design supplementing the upfront analysis findings.  For example during Phase C Information Systems Architecture - Application development, can be used to predict the impact on production systems of new or modified applications.

How effective is this?

TOGAF users at the London conference presented how simulation provided information assisting business decision makers select an 'optimised' IT solution based on the simulation data.
 

Simulation modelling options analysis is suitable for for business critical systems, mission critical systems and safety critical systems.  Simulation modelling can use specific organisation's goals and drivers as modelling criteria.  In the case of business critical systems such as banks or on line flight reservation systems -  the inability to trade and process transactions online can result in serious impacts to the organisations - both financially and reputationally.

TOGAF methodology identifies the capabilities required by the business,  capturing the essential drivers. The  Architecture Development Method is able to include analysis via simulation to evaluate  the infrastructure options.
 

In the example below simulations and modelling are applied to evaluate each option against the business goals and drivers that are deemed relevant in their organisation. 

The following criteria were used in the simulation modelling:

Manage Performance Risk.

8 scenarios of message flows through the target system based on capacity plan were used to evaluate performance risk based on the capacity plan.

Manage Availability Risk.

In order to assess managing availability risk the following goals were defined by the business.

Provide a highly-available (99.99%) and resilient services during the working day In case of database failure recover within 15 minutes; for data corruption within 2 hours.

18 scenarios of potential failures to test for availability were defined in order to analyse which options could provide a highly-available (99.99%) and resilient service during the working day. The simulation used the Monte-Carlo Analysis algorithm.

Inputs included: Mean time to recovery (MTTR), Mean time between failures, oracle availability and service availability. Outputs gave the percentage Service Availability figures.

Generally, non functional requirements tend to conflict and contradict each other( with the result that the design is a result of tradeoffs).

The analysis trade-offs considered Performance, Availability, Complexity, Usability, and Security
Findings were able to show:

Simulation modelling gave visibility of the following conflicts:
• Some critical complexity goals compromised to meet performance.
• Some non-critical Usability goals compromised to meet Complexity goals.

 How does this help the business make decisions related to the trade-off analysis?

Constraints, especially those concerned with human engineering issues are highly subjective and are best determined through complex empirical evaluations.

The data gives management visibility of the affected factors required to meet overall targets within optimised cost and risk boundaries specified by the business.  These factors are used as inputs in the simulation modelling.

Security can also be included in modelling scenarios. Security requirements and security scenarios highlight security vulnerabilities and communicate security needs. 

The simulations and modelling were used to inform management of the Total Cost of Ownership for the alternative solutions with acceptable Trade-Off.  The information supports management on their decision to select the most optimal solution.
 

BENEFITS?

Discussion of how beneficial the initial  iteration was to produce the report  reported that IT, operations and management found it useful, helping communication and collaboration between stakeholders. Evaluation of the findings after implementing showed that the initial iteration did not include all the necessary factors and criteria which affect cost options and capabilities. During complex transformation options that have multiple dependency and layers of costs this is not surprising. Setting stakeholder expectations on the costing granularity for the infrastructure options (which are still at the conceptual level) can help mitigate this.

Findings

- Use the TOGAF framework to capture the relevant  information from the various phases in a structured way.  This approach provides valuable input to the simulation modelling  (strategy, concerns and analysis of - where the organisation is now and where it wants to be etc.)  Simulation modelling provides valuable tradeoff analysis, reports and dashboard information improving visibility and clarity of the business landscape and enterprise architecture roadmap.

Simulation modelling depends on the quality of inputs - it can improve estimates, but is limited by the quality and availability of inputs.  Initial, high level modelling results will give rough orders of magnitude.  Subsequent information gleaned on subsequent design iterations, where more variables and data are available, will improve the granularity and accuracy. 

As with most techniques and tools knowledge of previous similar trade-off analysis within the industry helps to understand the variables involved and scenarios that are applicable.  Organisations within an industry may have varied drivers, so each organisation may require some variances to optimise solutions.

- Make it clear to all stakeholders that initial modelling is helpful to give a rough order of magnitude and further iterations rerun as the business transformation progresses can highlight additional detail and accuracy.  This helps set expectations and maintain trust.  It can also prompt the various stakeholder groups to take responsibility and become more proactive in including more details they feel are relevant during modelling.

- Emphasise that simulation modelling can improve as the business becomes more adept and knowledgeable on its use. Previous simulation etc can be available in knowledgebase's so that criteria, constraint and feedback are available for future work - thus improving accuracy and competency over time.

- Make it clear that typically, a model of the current state is constructed. This 'current state' model is tested and validated against historical data. Once the model is operating correctly, the simulation is altered to reflect the proposed capital investments. This 'future state' model is then stress-tested to ensure the alterations perform as desired. Make it clear if options have no historical data or current state data to support the analysis.

- The simulation success is dependent on the quality of the scenarios used to test the model. Incomplete scenarios can lead to untested paths and therefore unknown factors - whilst management may consider that an in depth analysis has reduced this uncertainty.

- Consider and seek out similar scenarios available which may be available within the industry or best practice libraries.

- Use existing knowledge and examples as starting points, and then modify/add organizational specific scenarios as required.

-Define and validate requirements and criteria with relevant stakeholder groups.  Validating and fine tuning the simulation model improves accuracy and knowledge base of relevant stakeholders who are included.

Conclusion
Helping business management make informed decisions and understand the various IT options open to them is undoubtedly beneficial, helping to improve communication between business and IT. It is a challenge to run simulations which factoring in all the various factors affecting the various option. Simulation can be used to evaluate transformation projects using defined goals, drivers and criteria.

The simulation success is dependent on the quality of the scenarios used to test the model. Incomplete scenarios can lead to untested paths and therefore unknown factors - whilst management may consider that an in depth analysis has reduced this uncertainty.

Also the measurement criteria used to rank the options is crucial to accuracy and outcomes - it helps if the criteria are known, clear and unambiguous. A clear understanding of effective measurement and scoring criteria is crucial. As with most tools application and usage is core - knowledge and experience will factor and influence in the quality of the outputs.

Organisations that apply the simulations effectively can benefit from informed options choices driving their capability roadmap forward.