SDITPM: a novel transfer policy model for agent server migration in e-business

Author

Lo, Jason C T ; Wong, Allan K Y ; Lin, Wilfred W K

Author_Institution

Dept. of Comput., Hong Kong Polytech., China

fYear

2005

Firstpage

581

Lastpage

584

Abstract

The novel statistical distribution independent transfer policy model (SDITPM) facilitates agent server migrations in an e-business setup that is based on pervasive computing. As a result it improves the serviceability of the setup because the nomadic users can happily get their service result within the period of their stay in the wireless smart space. The SDITPM works with proportional (P), derivative (D) and integral (I) control (i.e. PID approach). It can leverage many primary metrics (e.g. context switching cycle time and/or CPU utilization) from which the secondary metrics are derived for making transfer decisions by planar and vertical integrations.

Keywords

Internet; electronic commerce; software agents; statistical distributions; three-term control; ubiquitous computing; CPU utilization; PID approach; SDITPM; agent server migration; context switching cycle time; e-business; load balancing; pervasive computing; primary metric; proportional derivative integral control; secondary metric; statistical distribution independent transfer policy model; transfer decision; transfer probability; wireless smart space;

fLanguage

English

Publisher

ieee

Conference_Titel

Mobile Business, 2005. ICMB 2005. International Conference on

Print_ISBN

0-7695-2367-6

Type

conf

DOI

10.1109/ICMB.2005.88

Filename

1493665