a research agenda for accelerating adoption of emerging technologies in complex edge-to-enterprise...
TRANSCRIPT
A Research Agenda for Accelerating Adoption of Emerging Technologies in Complex Edge-to-
Enterprise Systems
Jay RamanathanRajiv Ramnath
Co-Directors, CERCS@OSU23rd April 2009
Progression to Edge-to-enterprise
1980s• Individual productivity adata
management
1990s • Business process adata interoperability
2000s• Business effectiveness ashared IT
services, business monitoring, business intelligence
2010s
• Business value – limited by (a) complexity of legacy applications and (b) scale change in problem scope
Complexity Challenge - ExampleNationwide
• Hurricane Katrina events caused high volumes and unexpected fluctuations in certain request types (claims)
• Customer service representatives needed to identify and triage the critical needs
• Request volumes caused change in application loads • There was impact across servers , the hardware and the
communication infrastructure• There were unexpected changes in performance due to IT
reallocations resulting in calls to the IT Help Desk• IT help desk now needed to know how to deal with these new IT
problems
All of the above are the multi-dimensional aspects of a single complex system • Technology by itself will not address all these aspects
Problem scope: Managing change from the edge through the enterprise
• Technology is not the biggest challenge with respect to adoption – Yesterday’s workshop showed ‘Cloud Computing’ is not just about IT and computing. – E.g. Developing SLA! To do it right requires investment in domain analysis, else the result is
conflict, and usually an over-provisioned, expensive infrastructure.• E.g. To leverage Clouds, you need:
– Computing tools and technology, – Economic analysis capabilities– Organizational change management– Business process reengineering
• Need to think evolution, not transformation – the paradigm is one of continuous measurement, management and improvement - at all dimensions of the business– E.g. for data center management need to understand more than the mechanics of
virtualization– Need to understand interactions between the facility, power and computing – Need a knowledge management process to support the evolution
Complexity in breadth rather than in depth should also drive the research agenda, Requires development of a single, integrated methodological framework
Framework Objectives
• Co-engineering of customer goals, business goals, operational goals, and technology goals
• Integration of creational, operational and evolutionary views of underlying components. Why?– Separation of functional and non-functional (application vs.
infrastructure, business transaction vs. IT transactions) aspects means (for example) we cannot easily correlate the network traffic to an application function and to business value (needed to argue that the network costs are appropriate!)
– Separation of creational, operational, and evolutionary aspects means (for example) disconnects in defining the impact of a change to an existing architecture
Next idea? Traceability-enabled Adaptive Complex Enterprise
First we need common abstractions and a shared theory, for example:An enterprise and its environment forms a complex system
consisting of a set of shared Agents that interact and are ‘interested’ in the value provided by others
An Agent is also a business value provider – is human or automated and autonomic (hides detail)
An Agent is also a customer/stakeholder interested in certain outcomes - Business, IT, Operations, Strategy- of other agents
All Agents can see the value of interaction with other agents, as authorized
Physical Agents are made visible through sensors
Common Abstraction a Enterprise Ontology Concept a Shared Agent
Customer
• satisfaction
• value
• location
Business
• cost
• efficiency
Operational
• service
• location
Infrastructure use
• time used
• throughput
Infrastructure
• Creation• Longevity• Evolution• LegacyInteraction
Adds Value
Agent Interaction Modeling
• Value is created (or not) when Agents interact• Provides the context for monitoring that provides
the traceability.– Identifies the linkages to instrument to get traceability
across layers – Helps develop policies and guidance for process, resource,
data use, security and assurance– Enables line of sight visibility into agent value for decision
making - e.g. what to charge for my service– Enables dynamic ‘collaboration’ between agents – they can
‘see and act’ accordingly
Example – Traceable Visualization of 311 Data of Interactions between Requests and Agents
Potential Improvement in Response Policies
System Dynamics- City helpdesk triage
example of interaction
throughput rates between multiple
roles (dynamic assignment to agents) using
Vensim
City services impact analysis identifying points of innovation
0 1 2 30
500
1000
1500
2000
2500
3000
3500
4000
4500
5000MobilitySecurityCommunication/ reponsibility delineationOnline web based facilities (ex e-payments)Budgeting/funding processQuality controlProcurement managementGeomapping/GIS/GPSDocument managementInteroperability/system integrationExtend 311 capabilitiesAsset/ Inventory managementMonitoring systemsDocument ImagingEnterprise tools/ process trainingBusiness process mappingTools/softwareWork order system managementOutage downtimesDocumentation of work productsAutomationEncryption servicesAccess ManagementMarketing servicesInfrastructure installation
Complexity
FTEs
Summary
Autonomic(Agent)
Traceable(Interactions in a set
of shared agents)
Adaptive(Behavioral change in
context)
Research Objectives
• A unified theory for Adaptive Complex Enterprise (ACE) systems – to replace silo-based experiential enterprise-related knowledge.– Enables tools for predictive management
• What happens when we move scope of services from individual to group to departments to enterprises to multi-enterprise?
• When, where and how are we really becoming more efficient?– Allows us to share principles and theories more effectively within the
community.
• Approach: Develop and validate theory through field experiments on real industry problems. – Needs adoption of ACE (at least for now) and Deep Industry-
University Collaboration