Event-based Microservices with Concurrency

Control and Dynamic Scaling using Message

queue and Kubernetes
Mohit Malviya(18BCS1575) Raj Kumar
Department of Computer Science and Engineering Assistant Professor, Dept. of CSE
Chandigarh University, Gharuan Chandigarh University, Gharuan
[email protected] [email protected]
Abstract— The rapid advances in network speed, reliability and setup manual and instructions for each and every service to
security create a growing need to move software and services make it better for anyone to pick up and operate on the
from local storage and processing to a user's machine to the modified service .Another benefit of this method is that it
management of cloud-accessible and highly scalable cloud
makes the maker of the application sharply focused on writing
services.The work presented in this paper proposes a novel
approach to solve problem with highly scalable system i.e
high level quality complex code as they do [3]. As that others
concurrency. In recent years clever and effective solutions have maintainers and developers will be watching on it. There are
been developed to solve such a problem but but scaling and also structural and design patterns that can help to distinguish
making it reliable and safe is still in progress. The work objects. i.e, you may have a arch where a service group
presented solves such a problem using async communication communicates via a central messaging provider or message
between microservices using message queues with publisher- queue.This queue contains a tons of messages from different
subscriber patterns by using open-source tool knows as NATS services provided in the architecture. Message queues or event
streaming and kubernetes and Node.js, with these tools the bus such as NATS Streaming are a good example.It often
presented work can be scaled for the concurrent users using a
happens over time that people start paying more attention to
method known as database versioning that is not inbuilt in any of
the services and tools mentioned above. this central bus and begin to know more about your domain.
Being very smart, that can be a error as creates difficulties to
make changes to the code that requires teams collaboration in
Keywords— Kubernetes (k8s), Node, NATS, Concurrency, different teams [3,4]. The general advice for those types of
Asynchronous (Async), Containers, Virtualization, architectures would be to keep them loosely coupled and let
Architecture(Arch)
them handle and mange the routing. Event based facilities are
I. INTRODUCTION good to work with in those situations.
A. Microservices architecture and design B. Need of Async Communication and Messaging Queues
Event Based Microservice architecture and design has Another way to communicate is message based or say
become the leading architecture and implementation style event based communication. The process is often used in
option in the software and technology driven companies. many Application infrastructure to make working of the
Microservice and its design is an architectural style that application more freely and more consistent. Message of event
emphasizes the division of the system into smaller and simpler based on messages is again the perfect decision to make the
resources that are deliberately designed to make the business app stronger, more efficient [5]. In the message based async
more integrated, and is the emergence of architectural style communication, often the usage queue vendor to deliver and
that focuses on traditional resources [1]. It takes this same maintain the message passed by the facilitator and it can be
approach and expands. it is loosely packed combined services continued if it is necessary; messaging seller shall verify the
that can be easily created, developed, and to be maintained passage of the message. With a messaging seller,
independently. Each of these services has a clear operational communication between services will be smooth. Generally,
roles and their different responsibility and they can there are two options for message based or event based
independently communicate with other resources through communication:
simple endpoints or their subscribers and publishers to solve a Point to Point Communication
large difficult business problem. Since the resources available The messaging queue or event bus shall be used for this
are small, they can be composed of one or many different communication. Message generating component of the
small groups from the start are separated by service roles and application, calls producer (sender), will send the message
responsibility [1]. which makes it easier to maximize the online to the seller of the same message and the service
effort to upgrade if needed.Once created and deployes, these interested in that message, called the buyer (recipient), will
services are loosely packed and also be utilized independently use Send message to that line and make other settings for that
of each other and that is why it is easy to identify hot message [6]. Once the messages or the event is sent by the
resources and measure them without working fully. event bus or messaging queue will only be absorbed by one of
Microservices also provides advanced error detection where in the recipient and report will be purged afterward use. If the
the case of an error inside in one service the whole recipient or the event interest is low, the event shall remain in
architecture of application does not stop to work. If the errors that message queue until the recipient starts their service up
are fixed, it can only be provided with the right module of and uses the events. For particular reason, message or event
service instead of re-using the whole application architecture. based on messages is the best options for strengthening the
Another advantage is the microservices architecture that microservice arch.
makes it easy to select the technical stack (programming
languages, database, etc) which is more suitable for required
Publisher-Subscriber
function (service) instead of having to a high-quality
For publisher-based communications in messages, the
approach, equivalent to a single size, has been developed [2].
message in the event vendor shall be used to store and retrieve
They are the ones who created, deployed and maintain and
the event or message is sent to the publisher-subscriber who
support it. There is no separate support teams or repair teams.
listens to the article will use the event. Unlike point to point
Other way to acquire the same goal is to have an open source
communications pattern, the event will be up for use to every
distribution internal model. By adapting this approach, an
subscribers that are listening and the title may or may not have
engineer who wants the changes in service can modify the
one or more than one subscribers. The event persists on the
application architecture, deploy the feature, and submit the
subject until we remove it [6].
pull request or merge request by himself instead of waiting for
There are also a few open source distribution vendors that can
the service maker or inchare to take and implement the
be used for message or event based on messages like Active-
necessary modified changes. For this architecture to function
MQ, Rabbit-MQ, Apache-Kafka etc. Each event vendor has
properly, correct operating documents are required with the
its advantages or disadvantages depending on the application
architecture requirements. Eg, Apache-Kafka has a speed of when you transfer your money between bank accounts when a
165000 per second over a single series and a high speed of student gets the bank balance when the money is taken out
3Million per second messaging but Apache-Kafka does not from the original bank account and before it is deposited into
support line-based communications and (point identification the recievers bank account, it appears that the money is lost in
pattern) especially topic based communication (publisher and the account. the bank process. Isolation is a site that assures
pattern registration). There are so many message vendor simultaneous access to data. Separation is used in a money
suitable for many project or application, each with its own management agreement. The easiest method is to have all the
pros-cons. We can choose a event dealer that matches and students waits the reader has finished, which is known as a
fulfill our requirements based on the features, good or bad read-in transaction. transactions are known for creating
In messaging-based communications, messaging apps, another tension between long term activities and performance reviews.
from line or topic, must know the pillars or arch of the Concurrency control with microservice aims it by keeping
standard message consumed or released by the manufacturer duplicate copies of each data object. By Which, each user gets
or publisher. what he wants to the site sees a database summary over a
period of time. Changes made by the author will not be
Most commonly used event communication is event-based noticed by other site users until the modifications have been
communication. Unlike conventional message based completed (or, in terms of the site: until the work is done.)
communication, the event based communication, especially in
event driven ,messaging providers don’t need to know the If the concurrency control arch needs to update data, it will
information of the message. With an event driven pattern, the not write over the real and absolute data object with newer
service simply appends the event to the subject in the event data, rather create a new version of the data object. So there
vendor and the resources that listens to that subjects will are many saved version. The visual version of each action
respond and communicate with each. depends on the level of separation used. The commonly
In addition to the event-driven pattern, event-based isolation rate used by concurrency control is abridgmentation.
communication is often used as an event finder in the CQRS By summarizing, the function views the data status as when
pattern [7], a pattern that creates a distinction between write the task began.
functionality (add / update / delete) and learning functionality
(select). The message vendor is often used as an event store MVCC provides timely reads. Read activity under
where we will apply the CQRS pattern to microservices. concurrency control usually using a timestamp with timezone
or ID to determine which DB status should be operation
C. Concurrency Control would be performed and read these versions of the data.
Concurrency means the ability to perform multiple tasks at Therefore operations are separated from each other without
once, which may increase the efficiency of the application. the need for a lock [8]. However, in addition to the
Tasks can be moved to run faster, at a given time, or to unnecessary locks, other MVCC details such as Oracle, are
respond to completing one or more tasks they rely used. Writes create a new version, while compatible reading
on.Concurrency can simply be said to perform multiple tasks accesses the older version.
at a time. It is necessary to increase the efficiency of time. If Keywords – MVCC(Multiversion concurrency control)
multiple operations attempt to access the same data, then D. Scaling using Kubernetes
incompatibility arises. Consistent cash management is Container based microservice arch have drastically created
required to keep the same data [7].For example, if we take the newer ways of development and deployment teams
ATM machines and do not use the corresponding currency, evaluate,deploy modern software. Container help businesses
many people are not able to withdraw money on time in to modernize by making it easier to measure and perform
different locations. That is where we need to agree. deployment of the applications, but containers also present
The problems that arise when not using concurrency are as newer complex challenges by building a completely new infra
follows - ecosystem [9].
Updates will be lost - Some actions make changes and some Software companies from startups to MNC now use many of
actions remove those changes. Some actions dismiss updates container cases every day, and that is the scale of the scale
of another function. they have to deal with.
Kubernetes-K8S, originally developed by Google, is an open
Unrestricted Reliance or Dirty Learning Problem - Flexibility container operating system designed to automate the
has been reviewed in a single transaction, at the same time management of installed applications. In fact Kubernetes-K8S
other transactions have begun and removed the value of the has established itself as a defacto level of container
variance where the variable can be revised or the commitment orchestration and is a prominent Cloud Native Computing
made at the beginning of the business false values or previous Foundation (CNCF) project, supported by key players such as
variable values this is a major problem. Google,Azure,Teradata,Apple and Meta .

Incompatible Recovery - Some processes update multiple

variables, some processes are in the process of updating that
variables, and the problem that arises is inconsistency of the
same variable in different situations.

Without concurrency, if one reads in a service at the same

time as some other else is writing, it is likely that the service
will deal a piece of data written in half or non consistent. E.g,
 method. In short, the review was initiated by taking the
followed steps in sequence:

Search Methods : The term “microservice”, “microservice”,

“micro-service” and “microservice architecture” are founded
in articles published in research papers, live tour conferences
and model building workshops. Three online libraries
IEExplore, ACM-DL and Scopus were used.

Selects of basic subjects: Prior to selection, articles were pre-

evaluated to relate to the particular questions related to this
formal mapping evaluation . The headings, abbreviations are
searched to any relevance and that is included or omitted for
the purpose of this study, based on the determining factor
After applying the conditions of discharge.

Keywords and segmentation: Once the papers have been

selected, quality assurance is performed to create a quality
work model. It was helpful to extract many possible estimates
of characters and categories.result was, the research method
used was found to be commonly used in this study and was
Figure 1. Kubernetes Architecture used to search papers such as: Exploratory
Study,Confirmation Study etc.
Kubernetes defines it easy to download and use applications
in microservice arch. It does this by building an invisible layer B. Indentifying keywords for event based microservices
over the applications team, so that devops teams can post their challenges
applications and allow Kubernetes to control the following
activities: Following the rounds of reviews, the following keywords
were founds that can be related to the challenged of event
 Managing the use of resources by application or group based microservices. Keywords are associated with the
 End load application distributed evenly across all hosting challenges identified.
infrastructure • Connectivity / Integration: Communication between services
 Automatically upload rating applications to all different and integrating it has many aspects to the design of
application situations microservices arch. Based a good strategy is important in
 Monitoring of resource usage and service restrictions to design. The strategy includes coming up with the appropriate
automatically stop apps from using multiple apps and protocol, request time expectations, expiration time and
restart apps endpoint design.
 Refer request from one service to another services if Keywords: Endpoint, REST endpoints, socket, TCP, API
there is a lack of resources on the service eg service dies gateway, proxy
 Automatic use of required resources made clearly • Event Service Delivery: In this various services are to be
available when a new host is added to the collection interact consistently. It is necessary for the services to have a
 Ease simplifying canary shipping and reversing regular and refined process where events can register and
perform themselves. This helps the listners to find storage
locations for other events. It involves determining the right
At Kubernetes, HorizontalPodAutoscaler automatically listeneing strategy and specifying how endpoints gates are
updates a workload application (such as Deployment or defined to report event availability.
StatefulSet), with the aim of automatically measuring
workload to suit the need. Horizontal scaling means that the
answer to extended uploads is to use multiple Pods [9] • Performance: It is often recognized that introducing small
resource structures in the product industry often adds
‘conversational’ connections between different events. Eg,
II. RESEARCH STRATEGIES
fulfilling a business activity requirement may result in
Map studies are evaulated complete and robust revisions of arranging multiple events calls together, which in turn
specific questions in a particular topic, aimed at identifying introduces additional ease of end-user experience. Due to
fillers in the textbooks and identifying where it can be better, limited circumstances, the data usually used by one
basic studies are needed to make them. In this paper,followed microservice is usually another.
is presented following the guidelines. Keywords: performance, speed, simulation.

A. Planning period • Faults and error resistance: Its system's poser to recover
From the start, lots of research questions were written for from failure. It is up to the microservices maintainers to
investigation whilst the study. The motive for the research consider that and provide appropriate ways to properly restore
question was revised and adjusted. Later, the selected papers or recover to any failure to slow down or relocate to other
were tested against a quality process and the regression services of the system. Its usually expected as cloudy areas
system is repeatedly come up with following the merging where Infra such as (IaaS) creates unavoidable failures.
Keywords: mistakes,error, failure, recovery,fault, tolerance.
• Safety: Security and hack resistance is a challenge that must Figure 3. Synchronous microservices architecture
be carefully considered in the design of microservices.
Services and events communicate with each other in a variety Callers can receive a response in just milliseconds or in a few
of ways that build trust relationships. In some systems, it is seconds. Despite the request delay, the caller is unable to
important for the user to be identified in all the network proceed to the next iteration till response from the server is
service chains that occur between microservice. O-Auth and received. REpresentaional state transfer – REST is an example
O-Auth-2 are solutions used by software architects to deal for sync microservice.
with security challenges.
Keywords: security, auth, O-Auth, O-Auth-2, reliability, An async microservice is in which it request for a resource
attack and the following response occur independently of each of the
outher service. A common practice of using async
• Tracking and Debugging: For small services programs, microservice is to use messaging queues or event brokers,
centralized tracking and debugging are essential for engineers such as Apache-Kafka or Rabbi-tMQ, to act as a event bus as
to get to know the behavior of the architecture as a whole. shown in Figure 4 below. One service will push a message to
converting monolithic systems to microservices uses common another service using the message queue or event bus [11,12].
solutions used to correct errors and profile systems. Various
strategies and techniques is to solve this problem. tracking is a
service capability to track a series of calls to recover single
activity or single user requests. debugging is another
important part of any services. Debug logs are important for
testing and debugging purposes. Special attention should be
given to carefully designing a centralized logging system and
integration system so that developers can continue to correct
errors appropriately.
Keywords: trace logs, login, sessions, debugging, profile
creation

• Application Performance Monitoring: APM is a sign of Figure 4.Asychronous microservice architecture

infrastructure. It involves measuring the performance of
services present in the arch.
Keywords: monitoring, performance, health monitoring Targeted service receives a timely message. Shipping service
is not locked for seller. It just explodes and forgets.
• Distribution services: Distribution and measurement services
are important aspects of infrastructure. However, choosing the A hybrid-microservice is one that qualifies consistent and
right service for right platform greatly influences the design harmonious interaction. Eg, the service will support both
of the microservices arch system. Container and its tools with HTTP- protocols and messages [12].
organized PaaS solutions offer a variety of features that make
posting and processing work pointless. However, choosing the
correct solution for service and syster arch is important as
every solution come with their own set of criteria and its own
ideas, which the architect must follow in order to be used to
use the chosen services in the best possible way. Measuring
event base microservice can be a challenge if the right arch is
not followed [10]. There must be several guidelines which has
to be followed to simplify service implementation and
recoveries. Many guidelines require flexibility and flexibility
in delivering service time to OS and on-site applications.
Keywords: tasks, containers, docker, rating, default scaling,
rolling feature updates, container images, replica-set

III. OUTCOMES AND INTERVENTION

A. Communication between services Figure 5.Hybrid microservices architecture
A synchronous microservice is where the data moves in
and out of the service end as if blocked interception. An As displayed in Figure 5 below, service B receives an request
example of synchronized message exchange is the HTTP from Service A. However, before Service B responds to
request-response interaction, which can be seen in Figure 3 Service A, it sends a message to Service C. The event sent to
below. If the request is made in under HTTP, the publisher is Service C will probably contain information relevant to the
locked in a session until a response from server is received concern of Service C.
[11].
Service B publish a message to Service C correspondingly,
without waiting a response from Service C. This method
pattern has a direct benefit to those affected by the speed of .
interaction.

GraphQL developed by Meta is a tech that supports hybrid

microservices. Callers can share a services published under
GraphQL with its synchronized queries and flexibility tech.
But the publisher can also receive the message equally from a
service using GraphQL subscriptions [14].

For this application we are using NATS Streaming a

interconnected technology designed for a highly
interconnected world. One technology that allows applications
to securely connect to a whole host of cloud, local, edge, web
and mobile, and devices. The architecture is as show below in
figure 6.

Figure 8. Kubernetes-K8S architecture

B. Containerization and scaling
The application used Horizontal pod autoscaler to scales the
Containerization is the process of containing together of pods horizontally means more servers containing the docker
working processed and code with all it's necessary image and uses cluster autoscaler to autoscale the worker
components like libraries, frameworks, and other nodes also uses an ingress controller to manage load balancing
dependencies in one self independent unit as suggested and api gateway services. Figure 8 explain the concept of
“containers” kubernetes and its components.
Today engineers use Docker to create modules called
microservices, which break down packages and break down
tasks into separate, stand-alone applications that work IV. RESULTS AND DISCUSSION
together. A. Microservices construction and architecture
Below is a diagram that define the architecture of docker a
container manager application suggestions for solutions and types of paper verification
studies were the main sources of answers to the question
because they pay more responsibility to structural architecture
than other papers. In particular, the architecture patterns and
design sections of those papers provided visual-diagrams used
and their through explanations. However, partial-contextual
architecture were used in the work, a huge type of other
graphic representations were also selected, although no
explicit reasons were given for the choice of a particular
architecture. This lack of persistence in graphic represerntaion
that may indicate the need to promote a more detailed view-
language modeling that better integrates and tells sub-services
Figure 7. Docker architecture based.

For scaling the application a tool knows as kubernetes is used The graphic ideas cited in the books was varied and could be
Kubernetes-K8S, originally developed by Google, is an open divided into random boxes , individual UML diagram sets that
container operating system designed to automate the include different layout features, vertical graphs and arrows
management of installed applications. In fact Kubernetes-K8S and finally SQL-NoSQL databases [15]. This means that the
has established itself as a defacto level of container microservice architecture process proposes to place a single
orchestration and is a prominent Cloud Native Computing microservice, i.e. a component, in a single functional area, that
Foundation (CNCF) project, supported by key players such as is a container, in order to create final independence and
Google,Azure,Teradata,Apple and Meta . separation of the microservice. In addition to the pre-covered
descriptive diagrams, descriptive programming languages are
also included in the paper to provide a more detailed view of
architectural details. The categories of different descriptive
language formats mentioned in the study include the
following:
• Common model data serialization languages, e.g. YAML
[16]
• Specially designed model languages,. [3] N. Alshuqayran, N. Ali and R. Evans, "A Systematic Mapping Study in
Microservice Architecture," 2016 IEEE 9th International Conference on
• Standard programming language specifications, e.g. Service-Oriented Computing and Applications (SOCA), 2016, pp. 44-51,
Javascript-Node.js, JSON and Typescript [16]. doi: 10.1109/SOCA.2016.15.
• Specific languages for special purposes. [4] F. Wan, X. Wu and Q. Zhang, "Chain-Oriented Load Balancing in
• Pseudocode algorithms [17,18]. Microservice System," 2020 World Conference on Computing and
Communication Technologies (WCCCT), 2020, pp. 10-14, doi:
10.1109/WCCCT49810.2020.9169996.
B. Quality attributes in relation with microservices [5] L. Abdollahi Vayghan, M. A. Saied, M. Toeroe and F. Khendek,
To address this question, it was initially necessary to make "Deploying Microservice Based Applications with Kubernetes:
Experiments and Lessons Learned," 2018 IEEE 11th International
standard pronouns, as many different words found on the Conference on Cloud Computing (CLOUD), 2018, pp. 970-973, doi:
pages showed the same meaning of the same attribute. Its 10.1109/CLOUD.2018.00148.
noted in the architecture that the service quality characteristics [6] L. Zhao et al., "Gridmedia: A Practical Peer-to-Peer Based Live Video
of event based microservice architecture such as modularity, Streaming System," 2005 IEEE 7th Workshop on Multimedia Signal
Processing, 2005, pp. 1-4, doi: 10.1109/MMSP.2005.248686.
scalability, autonomy and retention are cited and reviewed in
[7] R. Yilmaz and F. Buzluca, "A Fuzzy Quality Model to Measure the
almost the revised papers [19,20] For each quality attribute, Maintainability of Microservice Architectures," 2021 2nd International
we reviewed model drawings embedded in the same pages Informatics and Software Engineering Conference (IISEC), 2021, pp. 1-
that describe that quality. This crossroads tried to ask which 6, doi: 10.1109/IISEC54230.2021.9672417.
types of model is best suited to show and-or evaluate certain [8] K. Bakshi, "Microservices-based software architecture and
quality features in arch approaches," 2017 IEEE Aerospace Conference, 2017, pp. 1-8, doi:
10.1109/AERO.2017.7943959.
[9] P. D. Francesco, I. Malavolta and P. Lago, "Research on Architecting
Microservices: Trends, Focus, and Potential for Industrial
V. CONCLUSION Adoption," 2017 IEEE International Conference on Software
This systematic study carefully examines the lessons learned Architecture (ICSA), 2017, pp. 21-30, doi: 10.1109/ICSA.2017.24.
about microservices and relevant architectural challenges. to [10] G. Liu, B. Huang, Z. Liang, M. Qin, H. Zhou and Z. Li, "Microservices:
architecture, container, and challenges," 2020 IEEE 20th International
come up with some of the solutions and its plans for the Conference on Software Quality, Reliability and Security Companion
future. The research question investigates which architecture (QRS-C), 2020, pp. 629-635, doi: 10.1109/QRS-C51114.2020.00107.
patterns and ideas, in addition with any models, are used to [11] A. Koschel, I. Astrova and J. Dötterl, "Making the move to microservice
represent the structures of microservices. In addition, the final architecture," 2017 International Conference on Information Society (i-
research question, refers to the potential quality related to Society), 2017, pp. 74-79, doi: 10.23919/i-Society.2017.8354675.
event based microservice presented in the literature [12] R. Manciola Meloca, R. Ré and A. Luis Schwerz, "An Analysis of
Frameworks for Microservices," 2018 XLIV Latin American Computer
Event-driven microservice facilities provide a powerful, Conference (CLEI), 2018, pp. 542-551, doi: 10.1109/CLEI.2018.00071.
flexible, and well-defined way to solve business problems. [13] R. M. Munaf, J. Ahmed, F. Khakwani and T. Rana, "Microservices
The data connection structure allows universal access to Architecture: Challenges and Proposed Conceptual Design," 2019
important business events across the organization. Event International Conference on Communication Technologies (ComTech),
2019, pp. 82-87, doi: 10.1109/COMTECH.2019.8737831.
vendors meet this requirement very well, as they allow for
[14] L. Chen, "Microservices: Architecting for Continuous Delivery and
robust data editing, can deliver updates in real-time, and can DevOps," 2018 IEEE International Conference on Software
operate at large data rates. Data connections are always Architecture (ICSA), 2018, pp. 39-397, doi: 10.1109/ICSA.2018.00013.
strictly cut off from a dynamic business intelligence and use [15] M. Waseem and P. Liang, "Microservices Architecture in
it, freeing these needs from individual circumstances. This DevOps," 2017 24th Asia-Pacific Software Engineering Conference
distinction of concern allows the event vendor to continue to Workshops (APSECW), 2017, pp. 13-14, doi:
10.1109/APSECW.2017.18.
ignore the needs of the business mindset (without supporting
[16] G. Fu, Y. Zhang and G. Yu, "A Fair Comparison of Message Queuing
literacy), which in turn focuses on storing, storing, and Systems," in IEEE Access, vol. 9, pp. 421-432, 2021, doi:
distributing event data to consumers. 10.1109/ACCESS.2020.3046503.
[17] M. Y. Afanasev, Y. V. Fedosov, A. A. Krylova and S. A. Shorokhov,
"Performance evaluation of the message queue protocols to transfer
The data communication layer separates the ownership and binary JSON in a distributed CNC system," 2017 IEEE 15th
production of data from access to and use of it. Applications International Conference on Industrial Informatics (INDIN), 2017, pp.
no longer need to do duplicate work by providing an internal 357-362, doi: 10.1109/INDIN.2017.8104798.
business concept while providing ways to sync and direct [18] W. Wang, M. Ning and Y. Wang, "Software architecture based on
external access to other services. message queue," ISCTT 2021; 6th International Conference on
Information Science, Computer Technology and Transportation, 2021,
pp. 1-5.
Any service can use the firmness and resilience of an event [19] A. Tesliuk, S. Bobkov, V. Ilyin, A. Novikov, A. Poyda and V. Velikhov,
vendor to make its data more accessible, including those that "Kubernetes Container Orchestration as a Framework for Flexible and
use an event vendor to maintain its log status change. Effective Scientific Data Analysis," 2019 Ivannikov Ispras Open
Conference (ISPRAS), 2019, pp. 67-71, doi:
10.1109/ISPRAS47671.2019.00016.
[20] C. -C. Chang, S. -R. Yang, E. -H. Yeh, P. Lin and J. -Y. Jeng, "A
VI. REFRENCES Kubernetes-Based Monitoring Platform for Dynamic Cloud Resource
Provisioning," GLOBECOM 2017 - 2017 IEEE Global Communications
Conference, 2017, pp. 1-6, doi: 10.1109/GLOCOM.2017.8254046.
[1] P. Bak, R. Melamed, D. Moshkovich, Y. Nardi, H. Ship and A. Yaeli,
"Location and Context-Based Microservices for Mobile and Internet of
Things Workloads," 2015 IEEE International Conference on Mobile
Services, 2015, pp. 1-8, doi: 10.1109/MobServ.2015.11.
[2] G. Márquez and H. Astudillo, "Actual Use of Architectural Patterns in
Microservices-Based Open Source Projects," 2018 25th Asia-Pacific
Software Engineering Conference (APSEC), 2018, pp. 31-40, doi:
10.1109/APSEC.2018.00017.