Genetic Algorithm for Multi-Objective Optimization of Container Allocation in Cloud Architecture

Container virtualization is an emerging technology in cloud computing mainly due to its portability and lightweight features. Scheduling is a key task, performed by container management tool, which indirectly affects the characteristics of distributed software system in terms of availability, realizability, scalability, resources utilization, as well as power consumption. However, current schedulers only focus on some of the aforementioned aspects but not all. In this paper, a Many-Objective Genetic Algorithm Scheduler (MOGAS) is proposed to handle all such objectives to realize solutions with better characteristics. The proposed scheduler is compared with the Ant Colony Optimization (ACO)-based scheduler. Based on the proposed objective functions, simulation results show that MOGAS is better than the ACO scheduler in equally distributing tasks by 50%, assigning unique set of tasks per node by 40%, and reducing power consumption by 7%, on average.

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Optimizing scheduling decisions of container management tool using many‐objective genetic algorithm

Imdoukh

Ahmad

Alfailakawi

2019

“…18,19 It is considered as one of the advantages of GA. 18,19 It is considered as one of the advantages of GA.…”

Section: Introductionmentioning

confidence: 99%

“…In such systems, because of the repeated relations on various sites, the query optimization is very challenging. 18,19 It is considered as one of the advantages of GA. Therefore, in this paper, an Artificial Bee Colony Algorithm based on Genetic Operators (ABC-GO) is proposed to find a solution to join the query optimization problems in the distributed database systems.…”

mentioning

confidence: 99%

Join query optimization in the distributed database system using an artificial bee colony algorithm and genetic operators

Panahi

Navimipour

2019

As the main factor in the distributed database systems, query optimization is aimed at finding an optimal execution plan to reduce the runtime. In such systems, because of the repeated relations on various sites, the query optimization is very challenging. Moreover, the query optimization issue with large-scale distributed databases is an NP-hard problem. Therefore, in this paper, an Artificial Bee Colony Algorithm based on Genetic Operators (ABC-GO) is proposed to find a solution to join the query optimization problems in the distributed database systems. The ABC algorithm has the global-local search capabilities and genetic operators to create new candidate solutions for improving the performance of the ABC algorithm. The obtained results have shown that the cost of the query evaluation is minimized and the quality of Top-K query plans is improved for a given distributed query. Moreover, this method decreases the overhead.However, it needs a longer execution time. KEYWORDSartificial bee colony optimization, distributed database, genetic operator, query plan optimization INTRODUCTIONDue to the rapid growth of data in current systems, 1,2 a set of the interrelated databases, known as a distributed database, is stored on multiple computers to improve performance, reliability, accessibility, and modularity in such cases compared to the conventional centralized database system. 3-7 In today's world, every user of the system, whether an employee or a customer, needs access to the company's databases. The data in distributed databases may be repeated at different sites according to a distributed allocation plan. 8-10 As a result, an association among different sites is needed to create answers to a user query. 11 The query optimization is a critical subject for the distributed database management system. Therefore, the poor performance of database for answering the query is the result of poor selection of a query execution plan. 12The query is a statement or a group of statements that performs some simple database operations such as deletion, updating, and retrieving data. It plays a critical role in data management and retrieval. Commonly, the distributed queries are more complicated than centralized queries due to data distribution over different sites. 13 Consequently, the impact of efficient query processing is increasing in a large number of applications.A common method for identifying the Top-k objects is that all database objects are scored on some function. Therefore, the DQP (distributed query process) strategy aims to generate the query processing plans that reduce the amount of data transmission between sites and, as a result, reduces the response time of the distributed query. The query optimization is a critical subject for database management systems, 14 whose purpose is to define an appropriate execution plan for the user's query. 15 The cost of each plan should be produced due to more than one plan for such a query, which is significantly dependent on the amount of participation and data transfer betwe...

Nature‐inspired resource management and dynamic rescheduling of microservices in Cloud datacenters

Joseph

Chandrasekaran

2021

Distributed Cloud environments are now resorting to Cloud applications composed of heterogeneous microservices. Cloud service providers strive to provide high quality of service (QoS) and response time is one of the key QoS attributes for microservices. The dynamism of microservice ecosystems necessitates runtime adaptations and microservices rescheduling to avoid performance degradation. Existing works target rescheduling in hypervisor‐based systems, while ignoring the influence of configuration parameters of container‐based microservices. In an effort to address these challenges, this article describes a novel microservice rescheduling framework, throttling and interaction‐aware anticorrelated rescheduling for microservices, to proactively perform rescheduling activities whilst ensuring timely service responses. Based on periodic monitoring of the performance attributes, the framework schedules container migrations. Considering the exponentially large solution space, a metaheuristic approach based on multiverse optimization is developed to generate the near‐optimal mapping of microservices to the datacenter resources. Experimental results indicate that our framework provides superior performance with a reduction of up to 13.97% in the average response time, when compared with systems with no support for rescheduling.