Information Technology Reference
In-Depth Information
A Conceptual Framework
for Simulating Autonomic Cloud Markets
Simon Caton
1
,IvanBreskovic
2
, and Ivona Brandic
2
1
Karlsruhe Institute of Technology, Germany
simon.caton@kit.edu
2
Vienna University of Technology, Austria
{
breskovic,ivona
}
@infosys.tuwien.ac.at
Abstract.
One of the major challenges facing the Cloud paradigm is
the emergence of suitable economic platforms for the trading of Cloud
services. Today, many researchers investigate how specific Cloud mar-
ket platforms can be conceived and in some cases implemented. How-
ever, such endeavours consider only specific types of actors, business
models, or Cloud abstractions. We argue that market platforms for the
Cloud paradigm cannot (yet) be rigidly defined, and require the ability
to progress and evolve with the paradigm. In this paper, we discuss an
alternative approach: autonomic markets. Autonomic markets automat-
ically adapt to changed environmental conditions based upon a given
concept of “performance”. We describe the autonomic MAPE loop in
the context of electronic markets and consider the types of a knowledge
produced and required for decision making. Finally, we present a con-
ceptual framework for a market simulator, a critical tool for autonomic
markets, based upon experiences using the GridSim simulation tool.
Keywords:
Cloud Computing, Electronic Markets, Autonomic Com-
puting, Computational Economics, Market Simulation.
1
Introduction
Today electronic marketplaces are challenged by a highly dynamic context: high
product variability, unpredictable participant behaviour, and the emergence of
new actors as well as actor types. Consequently, market situations that have
previously been unimaginable will arise and novel theories and paradigms are
needed to facilitate and control them. Examples of new market contexts are
in the domains of Smart Grids, the Internet of Services and Cloud Computing
as well as Social and Collaborative Environments. Many of these domains are
already or will become inherently reliant upon the economic systems represented
by electronic markets that can address their allocation problems.
However, a key challenge is that we do not know the most fitting anatomy of
an appropriate market platform. Even assuming the existence of an “adequate”
platform, a subtle or disruptive change within the domain can mean that the
platform no longer satisfies its domain requirements. Today, electronic market
