Network Latency and Application Performance Aware Cluster Scheduling in Data Centers

Data-center-based cloud computing has revolutionized the way businesses use computing infrastructure. Instead of building their own data centers, companies rent computing resources and deploy their applications on cloud hardware. Providing customers with well-defined application performance guarantees is of paramount importance. A user's application performance is subject to the constraints of the resources it has been allocated and to the impact of the network conditions in the data center. Given the network latency variability observed in data centers, applications' performance is also determined by their placement within the data center. We present NoMora, a cluster scheduling architecture whose core is represented by a latency-driven, application-performance-aware cluster scheduling policy. The policy places the tasks of an application taking into account the expected performance based on the measured network latency between pairs of hosts in the data center. If a tenant's application experiences increased network latency, and thus lower application performance, their application may be migrated to a better placement. Experiments on a testbed and in simulations show that our architecture improves the overall average application performance by up to 32.5 and 42 percent, respectively, demonstrating that application performance can be improved by exploiting the relationship between network latency and application performance.