Mechanisms for AgilityDr David N AllsoppA presentation to ICCRTS 2006September 26-28, Cambridge, UK
3
Uncertainty and Change
• Military
• Political
• Commercial
• Social/Cultural
• Technical
ICCRTS 2006
4
A “Runtime World”
• A fixed threat -> a fixed defence?– out of touch with reality
• Design for a fixed zone -> limited to that zone– Adversary will push us out of that zone
• Open and “always-on” systems– More like medicine than engineering
ICCRTS 2006
5
Implications for C2 systems
• C2 systems need to be agile and flexible
• Reconfigurable at run-time
• Minimal built-in assumptions
• Greater effectiveness by keeping option space as open and large as possible
ICCRTS 2006
6
Flexibility?• Flexibility of system = integratability and modifiability of parts
• “Loose coupling”
• Attributes and behaviours include:– Not vendor- or platform-specific: use open standards– Components with clear, separate interfaces– Dynamic discovery of components– Composability– Hardware/address/location – independence– Failure as a normal condition
ICCRTS 2006
7
Limiting factor?
• Some believe that technology is the limiting factor or “rate-determining step” in agility
• We argue this is not the case:– We present a proof-of-concept software tool– We discuss technical mechanisms to achieve the
required flexible behaviours
ICCRTS 2006
8
Case Study: Decision Desktop
• Proof-of-concept agile information tool
• Acquire, visualise and manipulate diverse and dynamic battlespace information
• Create and share flexible visualisations to suit task
• Decision-makers can get:– the information they want– when they want it– in a form that suits the task at hand
ICCRTS 2006
9
Design philosophy
• Minimise design-time assumptions to maximise run-time flexibility
• Avoid arbitrary limitations on:– Where information comes from;– What information is required;– The level of detail of the information;– When information must be transferred;– Mode of information transfer (push or pull)– How the information will be displayed
ICCRTS 2006
Plug-in info sources
PalettePalette TimeDecoderTime
DecoderIcon
LibraryIcon
Library
Information
Context
LocationDecoder
LocationDecoder
Views:
Filter
Other plugins:
“Semantic cloud” of information
Filter
11
Technical Work
• Technical work
12
Integrating dataASW data in the Coalition Agents
Experiment
Agents & ServicesPlug-in
information resources
Integrating dataMTI tracks from Shared Tactical Ground Picture (STGP)
ComponentsPlug-ins for
selecting and displaying data
ContextShow areas of knowledge; use targeted queries
Drilldown Interact with
objects and drill down for more
detail
UncertaintyManage multiple
locations or uncertain status
Viewsconfigured to
display chosen information
13
Flexible visualisation
• Coalition Agents Experiment (2002)
• Rapid integration of data from new coalition partner
ICCRTS 2006
14
Key features
• Ability to handle diverse, unexpected data
• Plug in new functionality on-the-fly
• Tailor visualisations at runtime to suit task
• Share visualisations over network
• Not a static “picture” – objects’ data can be explored/browsed in a variety of views
• Plug-in filtering, symbology, data translation…
ICCRTS 2006
15
Discussion: Agility
• New demands on users? New roles?
• How will this form of agility be used?
• How does “bottom-up” agility interact with “top-down”reshaping?
• Can you have too much flexibility?
ICCRTS 2006
16
Discussion: ProcessesICCRTS 2006
Dynamic(run-time)
Fixed(design-
time)
Process-neutral
Process-centric
Decision Desktop
Office automation tools
Conventional C2 software
Agents, SOA?
17
How to get there?
• Technology is not the main obstacle– But C2 technology is not mature or complete!
• Don’t bend technology to fit old processes, then blame technology if little improvement is achieved
• How to design and procure technology that assists human creativity
• What tools and building blocks?
• Obstacles: specification, procurement, accreditation and exploitation of agile systems.
ICCRTS 2006
18
Acquisition
• Must deal with requirements change problem– Timescale of change != timescale of procurement
• Quality attributes such as flexibility – not easily measured– do not figure prominently in procurement requirements
• Evaluate systems not on what they can do one day one– What can they be modified to do in the future, and by
whom– Skills, licences, development tools, IP rights
ICCRTS 2006
19
Pointers for acquisition
• Reduce emphasis on specifying design-time requirements
• Approaches and metrics for assessing flexibility at design-time, build-time and run-time
• Embrace heterogeneity
• Exploit mix of architectures (e.g. Pub-sub, service-orientation)
• Bold experimentation; early deployment
ICCRTS 2006
20
Accreditation• Traditionally: off-line analysis of fixed configuration of a
stove-piped system– untenable in a network-centric design
• Complex open system -> emergent properties– Qualities of parts -> desirable behaviours of the whole
• New opportunities for security, as well as new challenges– Disciplined, compartmented architectures– Difficulty of attack
• D3C: run-time risk management / domain-based security
ICCRTS 2006
21
Conclusions
• We inhabit a “run-time” world of uncertainty and change.
• -> C2 systems need to be agile and flexible
• Improve agility by shifting from design-time “optimisation”towards run-time reconfiguration
• Such run-time tools are possible: Decision Desktop proof-of-concept
• -> Technology not the leading constraint – Experimentation, procurement, accreditation
ICCRTS 2006