Creating Decision Superiority: A Real-World Application of the Data Fabric

"Nothing is more difficult, and therefore more precious, than to be able to decide." Napoleon Bonaparte (1769 - 1821)

The French Emperor was recognised as a genius in the operational art of war. ?His innovation went beyond tactical deployment of troops.?Napoleon created corps, capable of fighting independently, and applied an early version (or so they say) of what would become the General Staff model as to how to conduct the analysis and planning necessary to deploy those same corps.????

I wonder what he would have thought of the direction from the Defence Strategic Review (DSR) 2023 to develop a “fully joined up and Integrated Force”.

I wonder also how he would have viewed the impact of the rapid translation of emerging and disruptive technologies into military capabilities.??And finally, I wonder how he would have demanded the integration of data and information.

There is perhaps no better reason to define and understand the practicalities and benefits of bridging reality and simulation than the need to make better decisions.?This is particularly important in the military context where decision superiority is vital – it wins wars and saves lives.???

Driving this challenge is the fact that our region “faces increasing competition that operates at multiple levels – economic, military, strategy and diplomatic- all interwoven and all framed by an intense contest of values and narratives”[1].?Australia must have a fully integrated and more capable ADF operating across five domains which work seamlessly together on joint operations to deliver enhanced and joined-up combat power”[2].??

We can define a high-level response in two layers:

1.????Experimentation to enable Force development.

2.????CoA analysis to conduct a campaign or a battle.?

I’ll use the term “commander” generically throughout:?In context the commander is a decision maker during both the conduct of operations and the creation of the Integrated Force deployed on those operations.?Their issues are the same – both must test options in a simulated environment with the right models and the right data.

Maintaining decision advantage is increasingly complex:??Defence is still operating under a decades old staff process that conducts manual analysis and evaluation of courses of action (the little Emperor may even recognise them), it lacks the ability to?display information in real-time, siloed tools make it difficult to provision and access modelling and simulation, and data and workflow inefficiencies emerge from the challenges of fusing large amounts of open and closed source data.

Whilst the technology has matured year on year and the focus today is about extracting the most value from the data, at the speed of the mission, it all boils down to one point: Defence suffers from a lack of intelligent information provided at the speed commanders need.?

A modern data architecture ensures data is accessible to relevant data users within their unique workflows. ?A Data Fabric is an architectural approach that simplifies data access in an organization and facilitates self-service data consumption. Teams can use this architecture to automate data discovery, governance, and consumption, through integrated end-to-end data management and governance capabilities.?They can then use data services to augment real data with synthetic data in modelling and simulation tools to evaluate courses of actions.??

IBM and Red Hat are collaborating with clients in individual nations and business partners to develop this into a service-based platform in line with?NATO’s Modelling & Simulation as a Service?concept.?

IBM’s Defence Simulation Analytical Service (DSaS) is an open source software platform built on?Red Hat OpenShift?that presents capabilities from an ecosystem of providers – both industry and defence – to analysts that can be assembled to support a wide range of scenarios.?OpenShift allows those capabilities to run on any cloud infrastructure provider, on premises and out to the edge.?This offers scale and global reach to support operations in contentious scenarios wherever they are in the world.???

At I/ITSEC, we demonstrated DSaS supporting a game for a scenario in Ukraine developed by Dr. Frederick W. Kagan of the?American Enterprise Institute?drawing on the work of the Russia?Team of the?Institute for the Study of War.?

The scenario recognises the importance of an additional information domain.?The information domain includes news, social media et al with origin and intent.?Having associated indicators and warnings is now essential to understanding the likelihood of offensive operations rather than relying only on traditional indicators in the physical domains.

DSaS presents information to analysts in various integrated views using artificial intelligence (AI) and machine learning (ML) to provide prioritised insights with confidence scores where appropriate and explainability of assertions back to their sources.?It does not prescribe full automation: analysts can explore what they see and data behind it, adjust indicators and warnings, and model scenarios.

Data sources are fed into a data fabric.?The data fabric uses IBM’s?Cloud Pak for Data?running on Red Hat OpenShift and provides information governance services and a catalogue to enable easy access to data for other tools.

For example, DSaS offers capability for live monitoring such as video analytics from cameras prioritised using AI.?Another example is that models can help detect misinformation and an analyst can teach the system to improve accuracy.?We combine the location and movement of physical assets with news, misinformation and cyber activity, just one way that we offer analysts a cross-domain perspective.?They can not only see the current situation but use timeline analysis for historical exploration and projection using AI.

Why is DSaS different???It allows commanders to make trusted decisions at the speed of relevance across the five domains highlighted by the DSR (air, land, maritime, space and cyber), based on, or including, open-source information.

A study by IBM’s Institute for Business Value highlighted an acceleration in adoption of AI in defence operations with information operations highest with 51% of departments piloting or implementing solutions.?Our work shows that combining information operations across the five domains is necessary to sustain superiority over adversaries.?

However, it is the area where defence needs most help.??The current suite of products processes information in office tools and evaluate courses of action manually using table-top exercises.?The few analytics tools that are deployed face difficulty accessing data, they are specialised and fragmented and therefore suboptimal for most tasks and nigh on impossible to integrate.?Furthermore, few command and control (C2) systems have analytics embedded.

I haven’t written much here on the importance of modelling and simulation and, for that matter, the need for partnerships and ecosystems to make this work.?I’ll draw out the key points at the https://www.simulationcongress.com/ in Adelaide 23 August.?Suffice to say, it has always been true that bad data creates bad outputs, regardless of the power and performance of the simulation suite.

The combination of a data fabric, analytics and AI, and modelling and simulation in the Defence Simulation Analytical Service (DSaS) offers trusted decision making at the speed of relevance.?

…the little General would be proud.

With thanks to Chris Nott?Global CTO for National Security,?Leendert van Bochoven?Global (including NATO) Leader for National Security,?and Stephen Gordon, DoD Account Lead for RedHat for the creation of the DSAS concept and its associated content.

[1] Defence Strategic Review (DSR) page 5

[2] DSR page 7