Coflow scheduling in input-queued switches: Optimal delay scaling and algorithms

Liang, Qingkai; Modiano, Eytan

doi:10.1109/infocom.2017.8057171

Cited by 18 publications

(3 citation statements)

References 35 publications

(61 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Because of the practical issues, a scheduling algorithm for stochastic real-time jobs in unreliable workers is crucial in distributed computing networks. The most relevant works to ours are [11,12]. While [11] focused on homogeneous stochastic jobs in the coflow model, [12] extended to a heterogeneous case.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Scheduling Stochastic Real-Time Jobs in Unreliable Workers

Hsu¹,

Huang²,

Shieh³

2019

Preprint

View full text Add to dashboard Cite

We consider a distributed computing network consisting of a master and multiple workers processing tasks of different types. The master is running multiple applications. Each application stochastically generates real-time jobs with a strict job deadline, where each job is a collection of tasks of some types specified by the application. A real-time job is completed only when all its tasks are completed by the corresponding workers within the deadline. Moreover, we consider unreliable workers, whose processing speeds are uncertain. Because of the limited processing abilities of the workers, an algorithm for scheduling the jobs in the workers is needed to maximize the average number of completed jobs for each application. The scheduling problem is not only critical but also practical in distributed computing networks. In this paper, we develop two scheduling algorithms, namely, a feasibility-optimal scheduling algorithm and an approximate scheduling algorithm. The feasibility-optimal scheduling algorithm can fulfill the largest region of applications' requirements for the average number of completed jobs. However, the feasibility-optimal scheduling algorithm suffers from high computational complexity when the number of applications is large. To address the issue, the approximate scheduling algorithm is proposed with a guaranteed approximation ratio in the worstcase scenario. The approximate scheduling algorithm is also validated in the average-case scenario via computer simulations.

show abstract

Section: Introductionmentioning

confidence: 99%

“…The most relevant works to ours are [11,12]. While [11] focused on homogeneous stochastic jobs in the coflow model, [12] extended to a heterogeneous case. The fundamental difference between those relevant works and ours is that we consider stochastic real-time jobs and unreliable workers.…”

Section: Introductionmentioning

confidence: 99%

Scheduling Stochastic Real-Time Jobs in Unreliable Workers

Hsu¹,

Huang²,

Shieh³

2019

Preprint

View full text Add to dashboard Cite

show abstract

“…Even for values as low as d = 2, these policies significantly outperform randomized splitting (which corresponds to d = 1) [4], [11]. In case of batch arrivals, the communication burden can be further amortized over multiple packets [15]. Power-of-d policies also extend to heterogeneous scenarios and loss systems (rather than singleserver queueing settings) [6]- [8].…”

Section: Introductionmentioning

confidence: 99%

Delay Versus Stickiness Violation Trade-Offs for Load Balancing in Large-Scale Data Centers

Liang

Borst

2017

2017 IEEE 25th International Symposium on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems (MASCOTS

Self Cite

View full text Add to dashboard Cite

Most load balancing techniques implemented in current data centers tend to rely on a mapping from packets to server IP addresses through a hash value calculated from the flow five-tuple. The hash calculation allows extremely fast packet forwarding and provides flow 'stickiness', meaning that all packets belonging to the same flow get dispatched to the same server. Unfortunately, such static hashing may not yield an optimal degree of load balancing, e.g. due to variations in server processing speeds or traffic patterns. On the other hand, dynamic schemes, such as the Join-the-Shortest-Queue (JSQ) scheme, provide a natural way to mitigate load imbalances, but at the expense of stickiness violation.In the present paper we examine the fundamental trade-off between stickiness violation and packet-level latency performance in large-scale data centers. We establish that stringent flow stickiness carries a significant performance penalty in terms of packet-level delay. Moreover, relaxing the stickiness requirement by a minuscule amount is highly effective in clipping the tail of the latency distribution. We further propose a bin-based load balancing scheme that achieves a good balance among scalability, stickiness violation and packet-level delay performance. Extensive simulation experiments corroborate the analytical results and validate the effectiveness of the bin-based load balancing scheme.

show abstract