Authors:
Jason C.T. Lo
;
Allan K.Y. Wong
and
Wilfred W.K. Lin
Affiliation:
Hong Kong Polytechnic Univeristy, China
Keyword(s):
SDITPM, primary metrics, secondary metrics, PID, planar integration, vertical integration, pervasive computing.
Related
Ontology
Subjects/Areas/Topics:
Mobile Software and Services
;
Telecommunications
;
Ubiquitous Wireless Services and Protocols
;
Wireless Information Networks and Systems
Abstract:
The novel statistical distribution independent transfer policy model (SDITPM) is proposed to improve the serviceability of a logical agent server in a pervasive computing environment. Serviceability is defined as the “chance of obtaining a required service within a defined period”. The SDITPM helps the agent to make sound migration decision by leveraging different primary metrics from which secondary ones are derived for the proportional (P), derivative (D), and integral (I) control elements. These elements are timely combined by planar and vertical integrations to form the final transfer probability that affirms a transfer policy decision to migrate. Therefore, The SDITPM is basically a PID controller that facilitates the decision making process.