Scheduling Storms and Streams in the Cloud

TL;DR

This paper introduces a randomized scheduling algorithm for streaming data processing jobs modeled as graphs, balancing load, cost, and queue lengths without preemptions in large server clusters.

Contribution

It presents a novel low-complexity, non-preemptive scheduling algorithm with a trade-off mechanism for minimizing partitioning cost and queue lengths in streaming data systems.

Findings

System remains stable under the proposed scheduling.

Algorithm achieves a smooth trade-off between cost and queue length.

No reliance on Gibbs sampler or preemptions.

Abstract

Motivated by emerging big streaming data processing paradigms (e.g., Twitter Storm, Streaming MapReduce), we investigate the problem of scheduling graphs over a large cluster of servers. Each graph is a job, where nodes represent compute tasks and edges indicate data-flows between these compute tasks. Jobs (graphs) arrive randomly over time, and upon completion, leave the system. When a job arrives, the scheduler needs to partition the graph and distribute it over the servers to satisfy load balancing and cost considerations. Specifically, neighboring compute tasks in the graph that are mapped to different servers incur load on the network; thus a mapping of the jobs among the servers incurs a cost that is proportional to the number of "broken edges". We propose a low complexity randomized scheduling algorithm that, without service preemptions, stabilizes the system with graph arrivals/departures; more importantly, it allows a smooth trade-off between minimizing average partitioning cost and average queue lengths. Interestingly, to avoid service preemptions, our approach does not rely on a Gibbs sampler; instead, we show that the corresponding limiting invariant measure has an interpretation stemming from a loss system.