Peer-2-Peer and Mobile Computing Models ❯ Adaptations to Enable the Thin Client Computing Model in Wireless Environments

Overview

Thin clients are uniquely suited for low-bandwidth networks and resource-poor devices. It is, therefore, attractive and opportune to exploit the thin client model for mobile and pervasive computing. However, wireless network latency and unpredictable application behavior could limit the utility of thin clients in the mobile environment. This research addresses optimization issues of thin clients in in the weakly-connected and intermittently-connected modes of communication.

Application-aware Adaptations

In this mode, we are investigating application-independent optimizations to preserve the advantage of thin clients under "active-media" behaviors. This includes applications with blinking cursor, blinking text, animated GIFs, repeated audio and other active components. We are also exploring application-specific optimizations aimed towards word-processing and keyboard interactions, in which user perceived response time is important. We use several approaches to realize this adaptation including detection and isolation of active components; detection and localization of keyboard input activities; and a wavelet based variable image compression techniques to enable rule-based adaptive compression by applications.

System-aware Adaptations

One approach we are investigating to cope with intermittent disconnection and to maintain QoS is the thin client based application adaptation. By switching computing modes between the proxy-based client/server model and the mobile thin client model, applications are enabled to participate in QoS assurance.

Objective

The overall self describing sensor network approach is based on the need to provide a scalable, plug and play, smart sensor network that is more flexible and maintainable. The approach leverages both mature and new technologies and standards to develop the system architecture composed of the sensor platform architecture, sensor network hierarchy, and associated software framework. The main approach consists of empowering the sensor network by using surrogate code located on the sensor node to allow interpretation of the data and sensor node control. Our sensor network hardware platform includes RF communications, EEPROM for data storage, EEPROM for surrogate storage, and microprocessor for communication and data manipulation.

Funding

Part of this research has been funded by CITRIX Systems Corporation. Organizations interested in sponsoring or collaborating to use these new technologies may contact us by email at helal@cise.ufl.edu.

People

Dr. Sumi Helal
Mohamed Al-Turkistany
Cumhur Aksoy
Siva Ramamurthy
Youngsup Kim

Publications

• Mohamed Al-Turkistany and Abdelsalam (Sumi) Helal, "Intelligent Adaptation Framwork for Wireless Thin-Client Environments," Submitted to the IEEE Symposium on Computers and Communications - ISCC'2003, to be held in KEMER - ANTALYA, TURKEY, June/July 2003 (pdf).

• C. Aksoy and A. Helal, "Optimizing Thin Clients for Wireless Active-media Applications," Proceedings of the Third IEEE Workshop on Mobile Computing Systems and Applications (WMCSA'00), held in Monterey, CA, Dec 7-8, 2000. (pdf).

• A. Ramamurthy and A. Helal, "Optimizing Thin Clients for Wireless Computing via Localization of Keyboard Activities," Submitted to the International Performance, Computing, and Communication Conference, to be held in Phoenix, Arizona, April 2001. (pdf).