Polaris Scheduler: SLO- and Topology-aware Microservices Scheduling at the Edge

Pusztai, Thomas; Nastić, Stefan; Morichetta, Andrea; Pujol, Víctor Casamayor; Raith, Philipp; Dustdar, Schahram; Vij, Deepak; Xiong, Ying; Zhang, Zhaobo

doi:10.1109/ucc56403.2022.00017

Cited by 11 publications

(3 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Then, we improve the algorithm with a randomized perturbation, specifically by modifying the perturbation node to the successor node of the current node in each iteration (lines 6-11). Subsequently, we use the Metropolis criterion to decide whether to accept the new solution obtained in each iteration (lines [12][13][14][15][16]. Finally, we reduce the overall temperature for cooling down, to enter the next iteration process (line 21), and output the optimal deployment set (line 22).…”

Section: Algorithm 1 Improved Simulated Annealing Algorithmmentioning

confidence: 99%

See 1 more Smart Citation

Microservice deployment strategy based on interaction weight in edge environment

He,

Xiong,

Zhang

et al. 2024

Preprint

View full text Add to dashboard Cite

This paper proposes a microservice deployment strategy based on interactive weights to optimize the deployment of microservices in edge environments. In an edge computing environment, this strategy is particularly important to reduce data transmission latency and improve processing efficiency as data is processed and stored closer to where the data is generated. The proposed deployment strategy evaluates the dependencies between microservices based on their interaction weights, and models computation and transport latency to minimize the overall latency of applications in the edge network. By introducing an interactive weight factor and constructing an end-to-end delay model of microservices, this strategy can effectively deal with the variability of microservice requests and the diversification of network transmission in edge computing environments. The experimental results show that this strategy is effective in reducing the overall latency of microservice deployment, and can deal with the interactions and dependencies between microservices more effectively than the traditional greedy algorithm (GA), random algorithm (RA) and other interactively aware deployment strategies (such as IntMA). In addition, experimental evaluations on open source microservices applications such as BookInfo validate the effectiveness of the proposed strategy.

show abstract

Section: Algorithm 1 Improved Simulated Annealing Algorithmmentioning

confidence: 99%

“…Typically, the invocation relationships between microservices are described as a Directed Acyclic Graph (DAG) [15]. Communication across multiple directed edges leads to interactions between nodes where microservices are deployed [16].…”

Section: Interaction Weightmentioning

confidence: 99%

Microservice deployment strategy based on interaction weight in edge environment

He,

Xiong,

Zhang

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Polaris scheduler is presented in [15]. It is an SLO-aware (Service Level Objective) scheduler that considers many network metrics.…”

Section: Related Workmentioning

confidence: 99%

Cost-Effective Scheduling for Kubernetes in the Edge-to-Cloud Continuum

Rac,

Brorsson

2023

2023 IEEE International Conference on Cloud Engineering (IC2E)

View full text Add to dashboard Cite

The edge to data center computing continuum is the aggregation of computing resources located anywhere between the network edge (e.g. close to 5G antennas), and servers in traditional data centers. Kubernetes is the de facto standard for container orchestration. It is very efficient in a data center environment, but it fails to give the same performance when adding edge resources. At the edge, resources are more limited, and networking conditions are changing over time.In this paper, we present a methodology that lowers the costs of running applications in the edge-to-cloud computing continuum. A cost-aware scheduler enables this optimization. We are also monitoring the Key Performance Indicators of the applications to ensure that cost optimizations do not impact negatively their Quality of Service. In addition, to ensure that performances are optimal even when users are moving, we introduce a background process that periodically checks if a better location is available for the application. To demonstrate the performance of our scheduling approach, we evaluate it on a vehicle cooperative perception use case, a representative 5G application.

show abstract