CMU-CS-02-113
Computer Science Department
School of Computer Science, Carnegie Mellon University



CMU-CS-02-113

Affinity Scheduling in Staged Server Architectures

Stavros Harizopoulos, Anastassia Ailamaki

March 2002

CMU-CS-02-113.ps
CMU-CS-02-113.pdf


Keywords: DBMS, databases, performance, staged server, scheduling, cache-conscious


Modern servers typically process request streams by assigning a worker thread to a request, and rely on a round robin policy for context-switching. Although this programming paradigm is intuitive, it is oblivious to the execution state and ignores each software module's affinity to the processor caches. As a result, resumed threads of execution suffer additional delays due to conflict and compulsory misses while populating the caches with their evicted working sets. Alternatively, the staged programming paradigm divides computation into stages and allows for stage-based (rather than request thread-based) cohort scheduling that improves module affinity.

This technical report introduces (a) four novel cohort scheduling techniques for staged software servers that follow a "production-line" model of operation, and (b) a mathematical framework to methodically quantify the performance trade-offs when using these techniques. Our Markov chain analysis of one of the scheduling techniques matches the simulation results. Using our model on a staged database server, we found that the proposed policies exploit data and instruction locality for a wide range of workload parameter values and outperform traditional techniques such as FCFS and processor-sharing. Consequently, our results justify the restructuring of a wide class of software servers to incorporate the staged programming paradigm.

22 pages


Return to: SCS Technical Report Collection
School of Computer Science homepage

This page maintained by reports@cs.cmu.edu