A dynamically configurable, multi-language execution platform

AbstractWe present a software platform for designing and testing wireless networks of sensors and actuators (WSNs). The platform consists of three components: an operating system for small-footprint microcontrollers (dubbed PicOS), a software development kit (SDK) amounting to a C-based, event-oriented (reactive) programming language, and a virtual execution platform (VUE2) capable of emulating complete deployment environments for WSNs and thus facilitating their rapid development.1 Its most recent incarnation introduced in the present paper is a component of the WSN lab being currently set up at Vistula in collaboration with Olsonet Communications Corporation.2 We highlight the platform’s most interesting features within the context of a production WSN installed at independent-living facilities.

Download Full-text

EcoGrid

Project Management ◽

10.4018/978-1-5225-0196-1.ch030 ◽

2016 ◽

pp. 607-623

Author(s):

Hemant Kumar Mehta

Keyword(s):

Resource Management ◽

Grid Computing ◽

State Of The Art ◽

Main Memory ◽

Statistical Distributions ◽

Memory Space ◽

Resource Management Algorithms ◽

Dynamically Configurable ◽

Management Algorithms

This chapter presents a toolkit for evaluation of resource management algorithms developed for Grid computing. This simulator named as EcoGrid and it is devised to support large number of resource or computing nodes and processes. Generally, grid simulators represent each resource using a thread that occupies large amount of space on the thread stack in main memory. However, EcoGrid models each node by an object instead of a thread. Memory space used by an object is much smaller than a thread, thus EcoGrid is highly scalable as compared to state-of-the-art simulators. EcoGrid is dynamically configurable and works with real as well as synthetic workloads. The simulator is bundled with a synthetic load generator that generates the workload using appropriate statistical distributions.

Download Full-text

EcoGrid

International Journal of Grid and High Performance Computing ◽

10.4018/ijghpc.2014070101 ◽

2014 ◽

Vol 6 (3) ◽

pp. 1-16

Author(s):

Hemant Kumar Mehta

Keyword(s):

Resource Management ◽

Grid Computing ◽

State Of The Art ◽

Main Memory ◽

Statistical Distributions ◽

Memory Space ◽

Resource Management Algorithms ◽

Dynamically Configurable ◽

Management Algorithms

This paper presents a toolkit for evaluation of resource management algorithms developed for Grid computing. This simulator named as EcoGrid and it is devised to support large number of resource or computing nodes and processes. Generally, grid simulators represent each resource using a thread that occupies large amount of space on the thread stack in main memory. However, EcoGrid models each node by an object instead of a thread. Memory space used by an object is much smaller than a thread, thus EcoGrid is highly scalable as compared to state-of-the-art simulators. EcoGrid is dynamically configurable and works with real as well as synthetic workloads. The simulator is bundled with a synthetic load generator that generates the workload using appropriate statistical distributions.

Download Full-text