To realize adaptation, applications first need to be aware of computing environment characteristics and their changes. In AwareWare, awareness is defined as the information of the computing environment characteristics. Awareness can be CPU utilization of a computing device, remaining battery power, available bandwidth between two end hosts on the network, or even an international variable of an application, e.g. the number of client connections at a server. AwareWare contains a set of tools that measure or monitor awareness information, and a management architecture to query, notify, and replicate the measured results. Our efforts in Awareness measurement and management focus on three aspects:

(1) integrating existing tools that are publicly available,

(2) developing new advanced tools that are more accurate and suitable for adaptive applications, and

(3) providing mechanisms for application developers to specify and customize these tools in the XML format.

 Besides intergrating exsisting tools, serveral new tools are developed. AwareWare¡¯s Java based bandwidth capacity measurement tool is discussed. The essence of the algorithm is a variation of the packet train probing model: to measure the bandwidth between nodes A and B, A sends a bunch of back-to-back packets to B, and B echoes back. Capacity bandwidth is estimated by infering the interval of the returned packets. The packet train is generated by a smooth traffic generator, which efficiently eliminates the side effect of traffic shaping in the network with ATM devices. It produces more accurate results and uses less probing packets compared to other tools.

The existence of wireless link introduces many new network dynamics for an distributed application, therefore AwareWare also includes a fuzzy reasoning algorithm to detect the existence of wireless links.

AwareWare also include an available bandwidth measurement tool, FEAT. In contrary to capacity measurement, available bandwidth refers to the minimum unused bandwidth of a path, which is generally smaller than the minimum link capacity along a communication path.

AwareWare also integrate wireless sensor networks (WSNs) to collect physical environmental information. A WSN consists of many tiny devices with sensing capability, (e.g. to measure temperature, or detect the location of a mobile device) and communicates with each another through wireless links. The ability to collect information from the physical environment can be useful for many adaptive applications, e.g. contextual computing based on location. AwareWare future includes mechanisms for over-the-air updating of the sensor codes (WSN reprogramming) to a WSN in its adaptation execution layer,