Professional Documents
Culture Documents
https://doi.org/10.1007/s13198-021-01195-8
ORIGINAL ARTICLE
Paola Ariza-Colpas1
Abstract Kubernetes is a portable, extensible, open-source individual terms has been greatly improved by the new
platform for managing containerized workloads and ser- system presented in this study. The KCSS presented in this
vices that facilitates both declarative configuration and study provides multicriteria node selection based on arti-
automation. This study presents Kubernetes Container ficial intelligence in terms of decision making systems
Scheduling Strategy (KCSS) based on Artificial Intelli- thereby giving the scheduler a broad picture of the cloud’s
gence (AI) that can assist in decision making to control the condition and the user’s requirements. AI Scheduler allows
scheduling and shifting of load to nodes. The aim is to users to easily make use of fractional Graphics Processing
improve the container’s schedule requested digitally from Units (GPUs), integer GPUs, and multiple-nodes of GPUs,
users to enhance the efficiency in scheduling and reduce for distributed training on Kubernetes.
cost. The constraints associated with the existing container
scheduling techniques which often assign a node to every Keywords Artificial intelligence Cloud infrastructure
new container based on a personal criterion by relying on Scheduling strategy Multi-criteria scheduler Automated
scheduling Kubernetes
123
Int J Syst Assur Eng Manag
occurs at the price of the implemented applications’ • Reduce the number of running containers.
specificity (Alicherry and Lakshman 2013; Ardagna et al. • Reduce the time it takes to send the user-selected item
2012). Because of portability and configuration, Kuber- towards its container.
netes is a common open source framework which is
designed for all infrastructure and services. It integrates
strategies including automated container optimization and
2 Proposed solution
distribution of specific nodes that are mainly dependent on
physical resource usages and specific threshold values, but
In this study, we presented KCSS, a new Kubernetes
it is commonly used among enterprises and academics.
Container Scheduling Technique for improving container
It is focused on a master–slave design, with the master
schedule efficiency in a mixed cloud computing environ-
and slaves communicating via a kubelet interface. The
ment. The objective is to address the complex issues that
unique feature of KCSS is how it includes a multi-criteria
organizations face implementing Kubernetes in a cloud
scheduling system for optimizing the allocation of a series
infrastructure and accepting online containers requested
of containers requested remotely through consumers. The
from its consumers, choosing the appropriate node to
purpose is to minimize the make span (the amount of time
manage every requested container based on multiple
it takes to organize and implement each container) as well
parameters relating to the user’s desire to decrease time to
as the average power efficiency. We suggest that hybrid
launch and the cloud provider’s willingness to reduce
parameters relevant to the environment for cloud comput-
energy usage. The presented solution suggests a load-bal-
ing and the customer must be included in our solution
ancing-based docker container task scheduling (Chinmay
(Ardagna et al. 2012).
and Roy 2012; Chinmay et al. 2011). The idea is that the
The artificial intelligence is playing a very significant
algorithm automatically manages the loading within each
role in scheduling of resources and decision making. For
container within a cluster threshold, ensuring that the
instance, the quality of Service (QoS) can be improved
loading is neither very heavy, neither very minimal for
using Artificial Intelligence (AI). With the advent of new
each container (Zheng et al. 2014; Baccarelli et al. 2015;
technologies and concepts including the Internet of Things
Beloglazov et al. 2012).
(IoT) and Artificial Intelligence, cloud computing became
The following are the key contributions of this work:
intelligent computation. Artificial Intelligence enables IoT
and Cloud nodes intelligent for their workload environment 1. Proposing KCSS as a new Kubernetes Container
and change in real time to support stronger QoS, lower Scheduling Technique.
power usage, and lower infrastructure costs (Sachin et al. 2. Take into account hybrid criteria in the KCSS, such as
2021; Amit et al. 2021, 2020). A scheduling technique for a the condition of the cloud infrastructure and the needs
cloud-based mutable environment, as well as specific of the users.
tasks are required. As a result, meta-heuristic-based cloud 3. In KCSS, find the node that has an equitable balancing
dynamic activity scheduling algorithms including Artificial of conflicting attributes for each container using a
Bee Colony (ABC), Ant Colony Optimization (ACO), and multi-criteria making decision method.
Particle Swarm Optimization (PSO) are preferred (Aroca 4. Integrate KCSS with Go language1 into the Kubernetes
et al. 2016; Elhady and Tawfeek 2015; Gill et al. 2019). system with minimal changes so that it can be used by
The following parameters were taken into account: (1) upcoming Kubernetes versions with ease.
the CPU usage level in every node; (2) the storage usage
Kubernetes is an open-source container improvisation
level of every node; (3) the disc usage level in every node;
tool that is widely used. As shown in Fig. 1, it automates
(4) the power usage within every node; and (5) the total of
the implementation, sizing, and maintenance of container-
operating containers throughout every node.
based applications. It is developed on the pods principle,
The major contribution of this study includes;
And Pod is made up of one or multiple containers that have
• Energy load assessment. been normally co-located, co-scheduled, and run through-
• Issues with load balancing. out the same way (Chen et al. Aug 2015; Chowdhury et al.
• Presented efficient load balancing and scheduling 2009; Cordeschi et al. 2012). As a consequence, these are
technique. all hosted by a virtual machine that already exists. It might
• CPU usage level assessment. be very well deploy on premises or in a public cloud. It
• Increase CPU usage. creates a Kubernetes cluster with a master and a set of
• Increase memory utilization. Kubernetes handlers. Scheduling in Kubernetes is broken
• Increase storage disc usage. down into two processes. The first process includes
• Reduce the power utilization. screening all nodes to eliminate those which may not
123
Int J Syst Assur Eng Manag
the lasting nodes applying priority (Dean and Ghemawat 2.1 Principle of the KCSS
2008). The priority is a key/value pair that represents the
priority’s name and the weighing selected from a list of The concept of KCSS is to pick the node that seems to have
options. The preference key function assigns a rating to an excellent balance around hybrid parameters linked to the
every node. Every priority function is given a positive structure of the cloud infrastructure and the consumer
value for each node’s total rating is determined from request for each recently requested container, as shown in
adding all of the weighted ratings together. Kubernetes Fig. 2. The specified node in our framework requires as
chooses the node with the greatest rating after all node follows:
scores have been computed (Dłaz et al. 2016; Fazio et al.
2016). 1. Increase CPU usage.
As shown in Fig. 2 the container request have been 2. Increase memory utilization.
generated by the users. It comes from Kubernetes 3. Increase storage disc usage.
scheduling for containers (KSC) method suggested by us. 4. Reduce the power utilization.
In regard to the KSC, then the resource controller can 5. Reduce the number of running containers.
manage the requests and shifts to the cloud infrastructure as 6. Reduce the time it takes to send the user-selected item
per the requirement and management of the resources to towards its container.
keep and continue the performance of the system (Felter We anticipate the five requirements mentioned above in
et al. 2015). KCSS. Nevertheless, other factors such as the cost of every
The following is a model for the problem analysis other node or the bandwidth length may be used to make
throughout this study. There is no no of nodes that make up the KCSS design more standardized. Each criteria recog-
the cloud infrastructure: node1 ; node2 ; and noden ;. For nized by KCSS is described in the following sections
this, there is a given set of c containers that would be (Guan et al. 2014).
submitted via the users for cloud services such as
container1 ; container2 and containern : The following will
be our tasks:
123
Int J Syst Assur Eng Manag
We assume that containers are requested online in our framework as mentioned in Fig. 1.
method. Such that, at time T0 , the scheduler has no idea Assume we have two nodes for following scenario at
how many and when containers are sent for next time T1 . time T0 :
123
Int J Syst Assur Eng Manag
123
Int J Syst Assur Eng Manag
123
Int J Syst Assur Eng Manag
Fig. 5 Ccontainer
implementation of node 1 with
Kubernetes approach
3 Results and discussion However, in the case of node 2, the average container has
more loads for scheduling and managing the system. We
As mentioned in Fig. 3 we have adopted a 2 node based result conclude that the large container with large resources
scenarios for energy scheduling and saving. The Fig. 5 have more capacity to manage the network and enhance the
depicts the working of node 1 for the scheduling of the performance of the network but need more time to get the
nodes with respect to the proposed scheduling technique. resources and execute. The outcomes of the simulation
While Fig. 6 presented the node 2 implementation results. revealed that the proposed technique using the AI concepts
123
Int J Syst Assur Eng Manag
Fig. 6 Ccontainer
implementation of node 1 with
Kubernetes approach
123
Int J Syst Assur Eng Manag
Chinmay C, Roy R, Pathak S, Chakrabarti S (2011) An optimal Guerrero C, Lera I, Juiz C (2018) Genetic algorithm for multi-
probabilistic traffic engineering scheme for heterogeneous objective optimization of container allocation in cloud architec-
networks. CIIT Int J Fuzzy Syst 3(2):35–39 ture. J Grid Comput 16(1):113–135
Chinmay C, Roy R (2012) Markov decision process based optimal Jiang J, Lu J, Zhang G, Long G (2013) Optimal cloud resource auto-
gateway selection algorithm. Int J Syst Algorithms Appl 48–52 scaling for web applications. In: International symposium on
Chowdhury N, Rahman M, Boutaba R (2009) Virtual network cluster, cloud and grid computing (CCGrid), pp. 58–65
embedding with coordinated node and link mapping. In: IEEE Kaewkasi C, Chuenmuneewong K (2017) Improvement of container
INFOCOM, pp. 783–791 scheduling for docker using ant colony optimization. In: 9th
Cordeschi N, Patriarca T, Baccarelli E (2012) Stochastic traffic International conference on knowledge and smart technology
engineering for realtime applications over wireless networks. (KST), pp. 254–259
J Netw Comput Appl 35(2):681–694 López-Torres S, López-Torres H, Rocha-Rocha J, Butt SA, Tariq MI,
Dean J, Ghemawat S (2008) Mapreduce: simplified data processing Collazos-Morales C, Piñeres-Espitia G (2019) IoT monitoring of
on large clusters. Commun ACM 51(1):107–113 water consumption for irrigation systems using SEMMA
Dłaz M, Martłn C, Rubio B (2016) State-of-the-art, challenges, and methodology. In: International conference on intelligent human
open issues in the integration of internet of things and cloud computer interaction (pp. 222–234). Springer, Cham
computing. J Netw Comput Appl 67:99–117 Onyema EM (2019) Integration of emerging technologies in teaching
Elhady GF, Tawfeek MA (2015) A comparative study into swarm and learning process in Nigeria: the challenges. Central Asian J
intelligence algorithms for dynamic tasks scheduling in cloud Math Theory Comput Sci 1(1):35–39
computing. In: 2015 IEEE Seventh international conference on Rimal Y, Pandit P, Gocchait S, Butt SA, Obaid AJ (1804) (2021)
intelligent computing and information systems (ICICIS) (pp. Hyperparameter determines the best learning curve on single,
362–369). IEEE multi-layer and deep neural network of student grade prediction
Fazio M, Celesti A, Ranjan R, Liu C, Chen L, Villari M (2016) Open of Pokhara University Nepal. J Phys Conf Ser 1:012054
issues in scheduling microservices in the cloud. IEEE Cloud Sachin D, Chinmay C, Jaroslav F, Rashmi G, Arun KR, Subhendu KP
Comput 3(5):81–88 (2021) SSII: Secured and high-quality Steganography using
Felter W, Ferreira A, Rajamony R, Rubio J, (2015) An updated Intelligent hybrid optimization algorithms for IoT. IEEE Access
performance comparison of virtual machines and linux contain- 9:1–16. https://doi.org/10.1109/ACCESS.2021.3089357
ers. In: 2015 IEEE International symposium on performance Shaheen Q, Shiraz M, Hashmi MU, Mahmood D, Akhtar R (2020) A
analysis of systems and software (ISPASS), pp. 171–172 lightweight location-aware fog framework (LAFF) for QoS in
Gill SS, Tuli S, Xu M, Singh I, Singh KV, Lindsay D, Tuli S, internet of things paradigm. Mobile Inf Syst. https://doi.org/10.
Smirnova D, Singh M, Jain U, Pervaiz H (2019) Transformative 1155/2020/8871976
effects of IoT, blockchain and artificial intelligence on cloud Zheng Y, Cai L, Huang S, WangZ (2014) VM scheduling strategies
computing: evolution, vision, trends and open challenges. based on artificial intelligence in Cloud Testing. In: 15th IEEE/
Internet Things 8:100118 ACIS international conference on software engineering, artificial
Guan X, Choi BY, Song S (2015) Energy efficient virtual network intelligence, networking and parallel/distributed computing
embedding for green dcs using dc topology and future migration. (SNPD) (pp. 1–7). IEEE
Comput Commun 69:50–59
Guan X, Choi BY, Song S (2014) Topology and migration-aware Publisher’s Note Springer Nature remains neutral with regard to
energy efficient virtual network embedding for green dcs. In: jurisdictional claims in published maps and institutional affiliations.
23rd International conference on computer communication and
networks (ICCCN). IEEE, pp. 1–8
123