On the theory of system administration

TL;DR

This paper develops a theoretical framework for system administration, integrating automation and human input, to optimize policies and strategies based on a model that predicts the best balance between resources and maintenance.

Contribution

It introduces a formal theoretical model for network and system administration, including an immunity model, to guide optimal policy implementation.

Findings

A mixture of automation and human input is necessary for effective system policy enforcement.

The immunity model predicts the optimal balance between resource availability and garbage collection.

Theoretical models can objectively relate policies to outcomes in system administration.

Abstract

This paper describes necessary elements for constructing theoretical models of network and system administration. Armed with a theoretical model it becomes possible to determine best practices and optimal strategies in a way which objectively relates policies and assumptions to results obtained. It is concluded that a mixture of automation and human, or other intelligent incursion is required to fully implement system policy with current technology. Some aspects of the author's immunity model for automated system administration are explained, as an example. A theoretical framework makes the prediction that the optimal balance between resource availability and garbage collection strategies is encompassed by the immunity model.