Service Level Agreements (SLAs) are currently one of the majorresearch topics in Grid Computing, as they serve as a foundation forreliable and predictable Grids. SLAs define an explicit statement ofexpectations and obligations in a business relationship betweenprovider and customer. Thus, SLAs should guarantee the desired anda-priori negotiated Quality of Service (QoS), which is a mandatoryprerequisite for the Next Generation Grids. The complexity of an SLA-based workflow grows significantly, as characteristics of correlated sub-jobs, the data transfer phases, the deadline constraints and possible failures have to be considered. Thus, an architect for an SLA-aware workflow implementation needs sophisticated mechanisms for specification and management, sub-job mapping, data transfer optimization and fault reaction. Therefore, this dissertation presents a system for SLA-aware Grid workflows. The main contributions include an improved specification language for SLA-aware workflows, three mapping and optimization algorithms for sub-job assignment to Grid resources, an error recovery mechanism, and a prototype implementation using standard middleware.