Lance Warren Russell

Dynamic Surge Protection: An Approach to Handling Unexpected Workload Surges with Resource Actions that Have Lead Times

Today’s information technology departments have significantly varying demands for resources due to unexpected surges in subscriber demands (e.g., a large response to a product promotion). Further complicating matters is that many resource actions done in response to surges (e.g., provisioning or de-provisioning an application server) have substantial delays (lead times) between initiating the resource action and its taking effect. This paper describes dynamic surge protection, an approach to handling unexpected workload surges in systems that have lead times for resource actions. Dynamic surge protection incorporates three technologies: adaptive short-term forecasting, on-line capacity planning, and configuration management. The paper includes empirical results from evaluations done on a research testbed, including favorable comparisons with a threshold-based heuristic. The results from an extended test also show that service objectives can be maintained cost-effectively.

Clockwork: A new movement in autonomic systems

Statically tuned computing systems may perform poorly when running time-varying workloads. Current work on autonomic tuning largely involves reactive autonomicity, based on feedback control. This paper identifies a new way of thinking about autonomic tuning, that is, predictive autonomicity, based on feedforward control. A general method, called Clockwork, for constructing predictive autonomic systems is proposed. The method is based on statistical modeling, tracking, and forecasting techniques borrowed from econometrics. Systems employing the method detect and subsequently forecast cyclic variations in load, estimate the impact on future performance, and use these data to self-tune, dynamically, in anticipation of need. The paper describes a prototype network-attached storage system that was built using Clockwork, demonstrating the feasibility of the method, and presents key performance measurements of the prototype, demonstrating the practicality of the methods.