Spindle

The Spindle effort focusses on issues in migration of applications to Infrastructure-as-a-Service clouds. Clouds may offer seemingly infinite resources, but scaling involves more than simply provisioning additional virtual machines (VM).

We construct performance models for applications based on profiling their constituent components during execution. We extract several features relating to CPU processing, memory consumption, and I/O. These performance models are continually updated to reflect the application's performance under conditions of varying load.

We then use these performance models to inform our VM placement and composition decisions while also reducing economic costs. To cope with tractability issues, we rely on using statistical and machine learning techniques to inform our algorithms that provide good performance under variable load conditions.

Key capabilties in the Spindle project include:

• Support for automated profiling of applications
• These applications can be multi-tier applications or streaming-based applications
• Autonomous scaling of multitier applications when performance thresholds are breached
• Deployments in real settings: private clouds and public clouds based on Amazon and Euclayptus.
• Interoperability with Xen and KVM hypervisors