TSpace

Abstract (summary): Due to the cost of sampling system performance, it is expensive to obtain performance characteristics of a complex computer system in different configurations. As an alternative, in this dissertation, we propose to reuse existing partial and full performance models and explicitly model the effects of configuration variations, with a goal of substantially reducing the time and effort required to perform performance reasoning and optimization on cloud-based systems, applications and services. We introduce Model Mapping, a novel inductive transfer learning technique for incremental performance modeling of highly configurable systems. Model Mapping captures many explicit and latent types of dynamic system evolution, including configuration changes, scaling and hardware upgrades, by deriving and modeling these kinds of incremental transformations between system and/or application instances, over time. Modeling these transformations allows us to build accurate models for new configuration instances with just a few samples.We experimentally test our method on a variety of system performance modeling and optimization scenarios, using a carefully designed experimental testbed and realistic benchmarks, to obtain insight on the method's applicability in real-world cloud computing environments. Among other examples, we show how our method can be used to quickly derive an accurate resource allocation split that optimizes a given overall performance goal for co-hosted applications in a virtualized environment. Compared to using conventional direct and incremental modeling techniques, our method achieves higher accuracy by up to an order of magnitude when the sampling budget is extremely limited, in particular when samples are limited to between 0% to 5% of an exhaustive sampling budget.