AUPY - Application Utility for Python Version: 1.0 Released on: November 20, 2007 Version: 2.0 Released on: October 10, 2008

==== IMPORTANT NOTICE:

The codebase for Aupy2 is outdated and needs to be improved. Some backwards compatible features (for Aupy1.0 style use) are buggy. A updated version, using Python2.7 and current best practices is being done in my spare time. The current version should still be considered proof-of-concept.

===== OVERVIEW:

AUPY is a bundle of utilties used to create and use utility functions within your python code to evaluate a program's performance and behavior. While the idea of application utility was originally developed for economic models and later applied in AI applications, it is extremely useful to create self-aware and self-monitoring applications; crucial in creating applications that exercise autonomic abilities and characteristics.

Autonomic applications exercise four core self-managing characteristics: Self-Configuration: Self-Healing: Self-Optimization: Self-Protecting:

Of these, AUPY can aid in achieving all except self-configuration.

These four characteristics can be employed at five levels of effectiveness within autonomic systems, as defined by IBM: Level 1 (Basic): All elements are managed independently; most systems today Level 2 (Managed): Systems are collected and managed in groups. System Awareness exists in some capacity and productivity is increased. Level 3 (Predictive): The system collects and monitors data, recognizing patterns of operation and behavior. This reduces the need for deep skills and system administration. Level 4 (Adaptive): The system uses the monitored data and recognized patterns to take specific actions known as Service-Level Agreements. These actions manage performance, healing, and protection. Level 5 (Autonomic): All components create a fully integrated and dynamic system completely managed by business rules and high level goals.

AUPY can aid in achieving up to level 4.

===== FILE MANIFEST:

aupe.py - all of the AUPY functionality and decorators. This file is used to define utility functions, functions to be monitored, atomic transaction functions, time-out functions, and more.

aupe_ftp.py - a self-aware FTP class for ftplib.

aupe_socket.py - a self-aware socket library. Currently it monitors throughput

binary_dl_client.py - an ftp client that is self-aware, using aupe_ftp, and optionally aupe_socket

example.py - example usage of aupe and its abilities

proc.py - a wrapper to gather information on linux systems from /proc

socket.py - a soft link to either aupe_socket.py or _org_socket.py, the originally socket library. Pointing this to aupe_socket.py inheritly makes ALL networked applications exercise the abilities of having application level utility for network operations.

===== NOTES:

AUPY IS DEPENDENT ON PCAP

There is still a lot of work to be done and aupy should be considered a proof-of-concept of bringing elementary level application utility into a language. I wrote the original functionality in about an hour to illustrate to the Software Engineering Research Group at Drexel University the benefits of basic autonomic abilities in software.

===== CONTACT:

Paul deGrandis [email protected]

Thanks for checking AUPY out!