Transfer Learning Based Performance Modeling And Effective Storage Management In Big Data Ecosystems

Big data ecosystems contain a mix of sophisticated hardware storage components to support heterogeneous workloads. Storage components and the workloads interact and affect each other; therefore, their relationship has to consider when modeling workloads or managing storage. Efficient workload modeling guides optimal storage management decisions, and the right decisions help guarantee the workload’s needs. The first part of this thesis focuses on workload modeling efficiency, and the second part focuses on cost-effective storage management.<div>Workload performance modeling is an essential step in management decisions. The standard modeling approach constructs the model based on a historical dataset collected from one set of setups (scenario). The standard modeling approach requires the model to be reconstructed from scratch with every time the setups changes. To address this issue, we propose a cross-scenario modeling approach that improves the workload’s performance classification accuracy by up to 78% through adopting the Transfer Learning (TL).<br></div><div>The storage system is the most crucial component of the big data ecosystem, where the workload’s execution process starts by fetching data from it and ends by storing data into it. Thus, the workload’s performance is directly affected by storage capability. To provide a high I/O performance in the ecosystems, Solid State Drive (SSD) are utilized as a tier or as a cache on big data distributed ecosystems. SSDs have a short lifespan that is affected by data size and the number of writing operations. Balancing performance requirements and SSD’s lifespan consumption is never easy, and it’s even harder when interacting with a huge amount of data and with heterogeneous I/O patterns. In this thesis, we analysis big data workloads I/O pattern impacts on SSD’s lifespan when SSD is used as a tier or as a cache. Then, we design a Hidden Markov Model (HMM) based I/O pattern controller that manages workload placement and guarantees cost-effective storage that enhances the workload performance by up to 60%, and improves SSD’s lifespan by up to 40%. </div><div>The designed transfer learning modeling approach and the storage management solutions improve workload modeling accuracy, and the quality of the storage management policies while the testing setup changes.<br></div>

Transfer Learning Based Performance Modeling And Effective Storage Management In Big Data Ecosystems

Big data ecosystems contain a mix of sophisticated hardware storage components to support heterogeneous workloads. Storage components and the workloads interact and affect each other; therefore, their relationship has to consider when modeling workloads or managing storage. Efficient workload modeling guides optimal storage management decisions, and the right decisions help guarantee the workload’s needs. The first part of this thesis focuses on workload modeling efficiency, and the second part focuses on cost-effective storage management.<div>Workload performance modeling is an essential step in management decisions. The standard modeling approach constructs the model based on a historical dataset collected from one set of setups (scenario). The standard modeling approach requires the model to be reconstructed from scratch with every time the setups changes. To address this issue, we propose a cross-scenario modeling approach that improves the workload’s performance classification accuracy by up to 78% through adopting the Transfer Learning (TL).<br></div><div>The storage system is the most crucial component of the big data ecosystem, where the workload’s execution process starts by fetching data from it and ends by storing data into it. Thus, the workload’s performance is directly affected by storage capability. To provide a high I/O performance in the ecosystems, Solid State Drive (SSD) are utilized as a tier or as a cache on big data distributed ecosystems. SSDs have a short lifespan that is affected by data size and the number of writing operations. Balancing performance requirements and SSD’s lifespan consumption is never easy, and it’s even harder when interacting with a huge amount of data and with heterogeneous I/O patterns. In this thesis, we analysis big data workloads I/O pattern impacts on SSD’s lifespan when SSD is used as a tier or as a cache. Then, we design a Hidden Markov Model (HMM) based I/O pattern controller that manages workload placement and guarantees cost-effective storage that enhances the workload performance by up to 60%, and improves SSD’s lifespan by up to 40%. </div><div>The designed transfer learning modeling approach and the storage management solutions improve workload modeling accuracy, and the quality of the storage management policies while the testing setup changes.<br></div>

Performance Modeling and Simulation for Water Distribution Networks

This study develops an implementation framework for asset management strategic planning of water distribution networks to meet sustainable infrastructure, socio-political, and financial targets over the life cycle of the infrastructure. The proposed framework is comprised of three decision-making layers: (1) Visions and Values, (2) Function, and (3) Performance. The asset management strategy framework is implemented and validated by demonstrating functionality and value by using data from three water utilities in Canada. The Visions and Values layer is set to meet the needs of the water utilities' stakeholders. The Function layer uses an advanced system dynamics model to simulate and forecast the system's future behavior. The Performance layer benchmarks, compares, and graphically illustrates the situation and performance of water utilities against each other regardless of their size. Benchmarking results indicate that all three water utilities can sustainably meet the strategic targets established in the Visions and Values layer of the asset management strategy over the benchmarking period. The impact of the desired cash reserve on infrastructure and financial benchmarking performance indicators is also investigated to explore the “optimal” combination of allowable fee-hike and rehabilitation rates using the contour plots developed over the benchmarking period. The results indicate that the optimal combinations of allowable fee-hike of ~8% per year and rehabilitation rate of 1.3% per year along with a 1–4% cash reserve, depends on the network condition, will allow water utilities to have sufficient funds to meet their strategic targets. The performance modeling and simulation approach presented in this study represents a powerful tool for other utilities to develop optimal strategic and operational plans for their networks and thus better service to their stakeholders.