Information Technology Reference
In-Depth Information
platforms are static and not conceived to handle changes in their domain or
elements of uncertainty in their architecture. Therefore, we argue that the market
engineering process cannot simply extend traditional approaches, but requires a
new methodology. In [6], we introduced an alternative: Autonomic Markets; a
goal orientated approach for market engineering to enable autonomic adaption.
To evaluate the vision of an autonomic market, we need an experimental
platform and it is not possible to simply map existing markets. Therefore, an
appropriate research methodology for their study is simulation, as it enables
the creation of what if scenarios and the ability to observe how autonomic
adaption evolves a market over time. Through simulation, we can implement
economic and management models of autonomic markets to access their perfor-
mance (with respect to goal fulfilment), tractability and feasibility for different
adaption strategies. Although real-life markets cannot be mapped directly for
an autonomic market, their traces as well as event and trading catalogues can
act as input data as a means to drive specific what-if scenarios. In this paper,
we reflect upon the lessons learned in [6], in order to create a set of requirements
and conceptual architecture for an autonomic market simulation tool.
This paper is structured as follows: Section 2, presents an overview of the
autonomic market vision; Section 3 discusses related work; Section 4 presents
a case study of GridSim in simulating markets; Section 5 draws upon this case
study articulate a conceptual framework for a autonomic market simulation tool;
finally Section 6 summarises the paper and discusses future work.
2 Autonomic Markets: An Overview
Our vision of an autonomic market platform is that
institutional forms
and
underlying
infrastructures
can be
adapted
at runtime with the goal of improv-
ing a given concept of
“market performance”
. Infrastructure adaption refers to
modifying the computational infrastucture of the market platform that enables
its core functionality. This, for example, includes computational resources, deliv-
ery mechanisms, communication channels, security procedures, etc. Institutional
adaption relates to modifying market components such as rules of participation,
allocation and pricing mechanisms, and tradable artefacts. Market performance
is characterised through specific goals that can include market liquidity, immedi-
acy, stability, security, participant welfare, energy eciency, allocation eciency,
etc. An institutional market form, i.e. an instantiated parameterisation of mar-
ket components, is what we refer to as a market configuration. Therefore, in-
frastructure adaption in our context is what we commonly understand as elastic
infrastructures in the Cloud paradigm and institutional adaption is a change in
one or more parameter settings and hence a change in market configuration.
By making market platforms autonomic, we hope to enable evolution beyond
initial design principles by “learning” or adapting towards ideal configurations,
possibly with certain levels of oscillation. Through this ability we can begin to
explore, analyse and evaluate autonomic market platforms as well as the impact
of different market configurations and goals. Autonomic adaption will enable
