ProMo: A Probabilistic Model for Dynamic Load-Balanced Scheduling of Data Flows in Cloud Systems

Souravlas, Stavros

doi:10.3390/electronics8090990

Cited by 14 publications

(11 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In our previous work, we implemented a CPN model for resource allocation policies in the cloud, where big data applications are executed [8]. This CPN model combined elements from a series of strategies on cloud systems and resource allocation, which we developed [9][10][11][12][13]. Other fields of PN and CPN applications are mentioned below: parallel processing [14], grid computing applications [15,16], traffic control [17], analysis of safety-critical interactive Systems [18], manufacturing [19], everyday applications [20], supply chain management [21], medicine [22], industry [23], project management [24], fuzzy systems [25], and communication protocols [26,27].…”

Section: Related Workmentioning

confidence: 99%

A Novel Method for General Hierarchical System Modeling via Colored Petri Nets Based on Transition Extractions from Real Datasets

2022

Self Cite

View full text Add to dashboard Cite

The Colored Petri net (CPN) has evolved into a complete modeling language, which is based on the object-oriented paradigm. Although this language is quite complete and can be used to accurately model a wide variety of systems, it seems that there is a lack of certain modeling rules, which can be used to generate models based on real datasets. This paper is a first attempt to incorporate sequences of real system events and translate them into sequences of Petri net transitions. Then, well-defined modeling rules control the model generation based on these transitions. The specific entities that take part in each of the real system events, as well as their attributes are also considered. The models produced using real data are structured because, in the majority of real systems, each incurring event usually affects a part of the system. Then, this effect is propagated to the entire system. Therefore, it is much easier to divide the entire model into distinct parts when real data are used and then interconnect these parts to build an entire model. This is a very important aspect when modeling very large systems. To test our approach, we used the real data from a courier company to generate its model. Our simulation results have shown that we managed to obtain quite accurate results through the model produced by the actual data.

show abstract

Section: Related Workmentioning

confidence: 99%

A Novel Method for General Hierarchical System Modeling via Colored Petri Nets Based on Transition Extractions from Real Datasets

2022

Self Cite

View full text Add to dashboard Cite

show abstract

“…From modular arithmetic, we are aware that, for linear Diophantine equations, a pair of processors (p, q) belongs to a communication class k if (pt − qs) mod g = k (9) Proposition 1 will make use of the definition of a class and Equation (7) to show the homogeneity of the processor pairs found in each class. Proposition 1.…”

Section: Mathematical Backgroundmentioning

confidence: 99%

“…In this work, we propose a pipelined modular arithmetic-based approach (the PMOD scheduler), which is based on the idea of each node receiving tuples for processing only from one other node at a time. The PMOD scheduler is proven to have some important advantages such as almost perfect load balancing, which is very important in today's Cloud systems [7], minimized buffer requirements and higher throughput. The PMOD organizes all the required operations (tuple transfer, tuple processing and tuple packing, which will be discussed in Section 5) in a pipeline fashion, thus decreasing the overall execution time compared to other known schemes, as the experimental results show.…”

Section: Introductionmentioning

confidence: 99%

Pipelined Dynamic Scheduling of Big Data Streams

Souravlas

Anastasiadou

2020

Applied Sciences

Self Cite

View full text Add to dashboard Cite

We are currently living in the big data era, in which it has become more necessary than ever to develop “smart” schedulers. It is common knowledge that the default Storm scheduler, as well as a large number of static schemes, has presented certain deficiencies. One of the most important of these deficiencies is the weakness in handling cases in which system changes occur. In such a scenario, some type of re-scheduling is necessary to keep the system working in the most efficient way. In this paper, we present a pipeline-based dynamic modular arithmetic-based scheduler (PMOD scheduler), which can be used to re-schedule the streams distributed among a set of nodes and their tasks, when the system parameters (number of tasks, executors or nodes) change. The PMOD scheduler organizes all the required operations in a pipeline scheme, thus reducing the overall processing time.

show abstract

“…Tang et al [37] introduced a dynamical load-balanced scheduling (DLBS) approach to improve the network throughput while balancing workload of data transmissions dynamically. Souravlas [38] addressed the problem of the balanced data flow among data centers. Tantalaki et al [39] introduced a pipeline-based linear scheduling approach for big data streams.…”

Section: Related Workmentioning

confidence: 99%

Minimizing Resource Waste in Heterogeneous Resource Allocation for Data Stream Processing on Clouds

Chung

Lee

et al. 2020

Applied Sciences

View full text Add to dashboard Cite

Resource allocation is vital for improving system performance in big data processing. The resource demand for various applications can be heterogeneous in cloud computing. Therefore, a resource gap occurs while some resource capacities are exhausted and other resource capacities on the same server are still available. This phenomenon is more apparent when the computing resources are more heterogeneous. Previous resource-allocation algorithms paid limited attention to this situation. When such an algorithm is applied to a server with heterogeneous resources, resource allocation may result in considerable resource wastage for the available but unused resources. To reduce resource wastage, a resource-allocation algorithm, called the minimizing resource gap (MRG) algorithm, for heterogeneous resources is proposed in this study. In MRG, the gap between resource usages for each server in cloud computing and the resource demands among various applications are considered. When an application is launched, MRG calculates resource usage and allocates resources to the server with the minimized usage gap to reduce the amount of available but unused resources. To demonstrate MRG performance, the MRG algorithm was implemented in Apache Spark. CPU- and memory-intensive applications were applied as benchmarks with different resource demands. Experimental results proved the superiority of the proposed MRG approach for improving the system utilization to reduce the overall completion time by up to 24.7% for heterogeneous servers in cloud computing.

show abstract

ProMo: A Probabilistic Model for Dynamic Load-Balanced Scheduling of Data Flows in Cloud Systems

Cited by 14 publications

References 28 publications

A Novel Method for General Hierarchical System Modeling via Colored Petri Nets Based on Transition Extractions from Real Datasets

A Novel Method for General Hierarchical System Modeling via Colored Petri Nets Based on Transition Extractions from Real Datasets

Pipelined Dynamic Scheduling of Big Data Streams

Minimizing Resource Waste in Heterogeneous Resource Allocation for Data Stream Processing on Clouds

Contact Info

Product

Resources

About