Resource Optimization Through Performance Modeling And Scheduling
Source: Silicon Valley Leadership Group
Synopsys' old data center environments had low utilization, averaging 10-20%, which resulted in excess use of hardware and energy. To improve, Synopsys centralized and then optimized access to its server and storage resources. Monitoring and scheduling optimize access to compute servers; defining Service Level Agreements (SLAs) helps users optimize use of compute and storage. To improve utilization of compute servers Synopsys adopted an aggressive management system involving monitoring and queuing for job control. It installed a monitoring tool on all compute servers in the data centers.