Microservices Management with the Unicorn Cloud Framework

George Feuerlicht

1,2 a

, Marek Beranek

and Vladimir Kovar

Unicorn College, V Kapslovně 2767/2, Prague 3, 130 00, Czech Republic

Prague University of Economics, W. Churchill Square. 4, 130 67 Prague 3, Czech Republic

Keywords: Cloud Computing, Frameworks, Microservices.

Abstract: The recent transition towards cloud computing from traditional on-premises systems and the extensive use of

mobile devices has created a situation where traditional architectures and software development frameworks

no longer support the requirements of modern enterprise applications. This rapidly evolving situation is

creating a demand for new frameworks that support the DevOps approach and facilitate continuous delivery

of cloud-based applications using microservices. In this paper, we first discuss the challenges that the

microservices architecture presents and then describe the Unicorn Cloud Framework designed specifically to

address the challenges of modern cloud-based applications.

1 INTRODUCTION

We are now in the midst of a major transformation

driven by wide adoption of cloud computing and

extensive use of mobile and IoT (Internet of Things)

devices. Public cloud platforms such as AWS

(Amazon.com, 2017), Microsoft Azure (2017), etc.

offer ever increasing range of elastic services, and

practically unlimited compute power and storage

capacity, allowing more flexible acquisition of IT

resources and avoiding most of the limitations of

traditional on-premises IT solutions. This new

technology environment is creating opportunities for

innovative solutions at a fraction of the cost of

traditional on-premises enterprise applications. To

take full advantage of these developments,

organizations involved in the production of enterprise

applications must adopt a suitable enterprise

architecture and application development

frameworks. Two key architectural requirements

emerge as a result of recent technological advances:

1) the requirement to develop and deploy server-side

application components on a cloud infrastructure in

the form of microservices, and 2) the need to support

various types of mobile and IoT devices using cross-

platform client-side application components.

Deploying application components on a cloud

infrastructure enables applications to be shared by

https://orcid.org/0000-0001-9333-5050

https://orcid.org/0000-0003-0491-4275

very large user populations. The architecture needs to

facilitate rapid incremental development of

application components, ensure secure access to

information and support rapid cloud deployment.

There is increasing empirical evidence that to

effectively address such requirements, the

architecture needs to support microservices and

container-based virtualization (Balalaie et al., 2016).

However, it is also becoming clear that the

management of large-scale clusters of containers has

its own challenges and requires automation of

application deployment, auto-scaling and effective

control of resource usage. The need for continuous

delivery and monitoring of application components

impacts on the composition and skills profile of IT

teams, favoring small cross-functional teams and

leading to convergence of development and

operations (DevOps). The separation of code

development and declarative methods of environment

configuration play an important role in increasing the

productivity of the software development process.

Furthermore, developers of enterprise applications

are increasingly turning towards open source

solutions that allow full control over the entire

software stack, avoiding costly proprietary solutions.

Also, while the use of public cloud platforms is

economically compelling, an important function of

the architecture is to ensure independence from

826

Feuerlicht, G., Beranek, M. and Kovar, V.

Microservices Management with the Unicorn Cloud Framework.

DOI: 10.5220/0009572008260832

In Proceedings of the 22nd International Conference on Enterprise Information Systems (ICEIS 2020) - Volume 2, pages 826-832

ISBN: 978-989-758-423-7

individual cloud providers, avoiding a provider lock-

in. Finally, the architecture should reduce the

complexity of application development and

maintenance process and facilitate effective reuse of

application components and infrastructure services.

These requirements demand a revision of existing

architectural principles and application development

methods.

In this paper, we describe the Unicorn Cloud

Framework (uuCloud). uuCloud is a part of the

Unicorn Application Framework (UAF), a suite of

closely integrated frameworks described in previous

publications (Beranek et al., 2017), (Beránek et al.,

2017), (Beranek et al., 2018). UAF was developed by

Unicorn (https://unicorn.com/cz) - a provider of IT

solutions based in Prague, Czech Republic and is used

to develop solutions for organizations across a range

of industry domains, including banking, insurance,

energy and utilities, telecommunications,

manufacturing and trade. The main objective of

uuCloud framework is to document and automate the

development and deployment of microservices

applications and to facilitate reuse of skills and

software components across multiple projects.

The paper is structured as follows. In the next

section (Section 2) we review recent research

publications dealing with frameworks for container-

based microservices. The following section (Section

3) describes the uuCloud framework and Section 4

are our conclusions.

2 RELATED WORK

In this section we review recent research publications

dealing with frameworks for the management of

container-based microservices. Cloud-based

application development frameworks and

architectures have been the subject of intense recent

interest by both industry practitioners and academic

researchers, in particular in the context of

microservices and DevOps (Mahmood and Saeed,

2013), (Thönes, 2015). According to Rimal et al.

(Rimal et al., 2011) the most important current

challenge is the lack of a standard architectural

approach for cloud computing. The authors explore

and classify architectural characteristics of cloud

computing and identify several architectural features

that play a major role in the adoption of cloud

computing. The paper provides guidelines for

software architects for developing cloud

architectures.

While container technologies and microservices

have revolutionized application development and

deployment, it is also evident that the use of these

technologies has its challenges. The management of

large-scale container-based environments requires

automation to ensure fast and predictable application

deployment, auto-scaling and control of resource

usage. At the same time, there is a requirement for

portability across different public and private clouds.

To address such issues a number of open source

projects aiming to support the development and

deployment of cloud-based applications have been

recently initiated. Prominent examples include Cloud

Foundry (CloudFoundry, 2017), OpenShift

(OpenShift, 2020), Docker Swarm (Naik, 2016) and

Kubernetes (Burns et al., 2016b). A key idea of these

open source platforms is to abstract the complexity of

the underlying cloud infrastructure and present a

well-designed API (Application Programming

Interface) that simplifies the management of

container-based cloud environments. Brewer

(Brewer, 2015) in his keynote “Kubernetes The Path

to Cloud Native” argues that we are in middle of a

great transition toward cloud native applications

characterized by highly available unlimited

“ethereal” cloud resources. This environment consists

of co-designed, but independent microservices and

APIs, abstracting away from machines and operating

systems. The Kubernetes project initiated by Google

in 2014 as an open source cluster manager for Docker

containers has its origins in an earlier Google

container management systems called Borg (Burns et

al., 2016b) and Omega (Schwarzkopf et al., 2013).

Kubernetes objective is to facilitate cloud native

systems that run applications and processes in

isolated units of application deployment (i.e. software

containers). Containers implement microservices

which are dynamically managed to maximize

resource utilization and minimize the cost associated

with maintenance and operations. The state of objects

in Kubernetes is accessed through a domain-specific

REST API that supports versioning, validation,

authentication and authorization for a diverse range

of clients (Burns et al., 2016a). Other similar efforts

in this area include OpenStack (Kumar et al., 2014) -

a cloud operating system designed to control large

pools of compute, storage, and networking resources,

managed using a dashboard used by administrators to

control cloud resources and to provision resources

through a web interface. In (Feller et al., 2012), the

authors argue that many existing frameworks have a

high degree of centralization and do not tolerate

system component failures, and present the design,

implementation and evaluation of a scalable and

autonomic virtual machine (VM) management

framework called Snooze. Truyen et al. (Truyen et al.,

Microservices Management with the Unicorn Cloud Framework

827

2019) evaluate the performance overheads introduced

by Docker engine and two container orchestration

frameworks: Docker Swarm and Kubernetes, when

running CPU-bound workloads in OpenStack. The

authors report that Docker engine deployments

running in host mode exhibit negligible performance

overheads in comparison to native OpenStack

deployments, but that virtual IP networking

introduces a substantial overhead in Docker Swarm

and Kubernetes. They conclude that the solution

involves service networking approaches that run in

true host mode and offer support for network isolation

between containers. Another finding discussed in the

paper was that volume plugins for persistent storage

have a large impact on the overall resource model of

a database workload. The paper provides

experimental confirmation for this assertion showing

that a CPU-bound Cassandra workload changes into

an I/O-bound workload in both Docker Swarm and

Kubernetes. This implies that placement decisions for

native or Docker engine deployments need to be

recomputed when using container orchestration

frameworks such as Docker Swarm or Kubernetes.

The authors conclude that while container

orchestration frameworks provide various advantages

for automating SLO-aware placement of multi-tenant

applications, to gain full benefits of this approach

container orchestration frameworks need further

development.

While Kubernetes is gaining momentum as a

platform for the management of cloud resources with

support from major public cloud platforms including

Google Cloud Platform, Microsoft Azure and most

recently Amazon AWS, there is a number of

alternative projects that aim to address the need for

universal framework for the development and

deployment of cloud applications. However, some

proposals lack empirical verification using large-

scale real-life applications. Complicating this

situation is the rapid rate of innovation in this area

characterized by the current trend toward finer

resource granularity with corresponding impact on

the complexity of cloud resource management

frameworks (Baldini et al., 2017). According to Ben-

Yehuda (Agmon Ben-Yehuda et al., 2014), the trend

toward increasingly finer resource granularity is set to

continue resulting in improved flexibility and

efficiency of cloud-based solutions. The authors

assert that resources such as compute, memory, I/O,

storage, etc. will be charged for in dynamically

changing amounts, not in fixed bundles. They

conclude that the RaaS (Resource as a Service) cloud

requires new mechanisms for allocating, metering,

charging for, reclaiming, and redistributing CPU,

memory, and I/O resources among multiple clients

every few seconds.

3 uuCLOUD FRAMEWORK

As noted in section 2, currently there is a lack of

agreement about a standard framework for the

management of cloud-based microservices. Our

solution is the uuCloud cloud management platform

that facilitates provisioning, monitoring and

management of elastic cloud services in the form of

virtualized Unicorn Applications (uuApps). uuCloud

supports hybrid cloud operation combining on-

premises cloud infrastructure with public cloud

infrastructures such as Microsoft Azure and AWS

cloud platforms.

Importantly, uuCloud enables the development of

cloud-native applications using the DevOps

approach. As illustrated in Figure 1, uuCloud

incorporates a comprehensive range of infrastructure

services that allow rapid deployment of applications

without the need to re-implement basic functionality

(i.e. authentication, authorization, etc.) separately for

every application.

Figure 1: uuCloud framework services.

Logging and monitoring services are used by

systems administrators at runtime to ensure that the

system operates as specified by the SLA (Service

Level Agreement).

The uuCloud framework consist of three basic

components: 1) uuCloud Operation Registry – a

database that maintains active information about

uuCloud objects, 2) uuCloud Control Centre (uuC3) -

a tool that is used to automate deployment and

operation of container-based microservices, and 3)

Operation Runtime that provides facilities for

ICEIS 2020 - 22nd International Conference on Enterprise Information Systems

828

monitoring runtime services. To ensure portability

and to reduce overheads uuCloud uses Docker

container-based virtualization (Docker, 2015).

Docker containers can be deployed either to a public

cloud infrastructure (e.g. AWS or Microsoft Azure)

or to a private (on-premises) infrastructure (e.g.

Plus4U – Unicorn cloud infrastructure).

uuCloud supports multi-tenancy and can

accommodate a large number of Tenants using

scalable resources across multiple cloud platforms.

Each Tenant typically represents a separate

organization (e.g. a publisher of online learning

courses). Tenants are assigned resources from

Resource Pools using the mechanism of Resource

Lease that specifies usage constraints such as the

maximum number of vCPUs (virtual CPUs), RAM

size and size of storage that are available for the

operation of a specific Tenant. A Tenant can be

allocated several Resource Pools, for example

separate Resource Pools for production and

development. Each Tenant consumes its own

resources and is monitored and billed separately.

Applications can be shared among multiple Tenants

with the Tenant in whose Resource Pool the

application is deployed being responsible for the

consumed resources. Each Host (physical or virtual

server) is allocated to a logical aggregation of

computing resources called a Resource Group.

Resource groups can have additional Resources (e.g.

MongoDB (MongoDB, 2020), Oracle DBMS,

Gateways, etc.). Each Resource Group belongs to a

Region – a separate IT infrastructure sourced from a

single cloud provider with co-located compute,

storage and network resources characterized by low

latency and high bandwidth connectivity. Regions are

grouped into Clouds that can be composed of multiple

public and private cloud platforms.

A Node is a unit of deployment with hardware

characteristics that include virtual CPU count

(vCPU), RAM size and ephemeral storage. Nodes are

classified according to NodeSize, e.g. M (Medium

size: 1xvCPU, 1 GB of RAM, 1 GB of ephemeral

storage) or L (Large size: 2xvCPU, 2 GB of RAM, 1

GB of ephemeral storage). Nodes are further

classified as synchronous or asynchronous depending

on the behaviour of the application that the node

virtualizes. Nodes are grouped into sets of nodes with

identical functionality called NodeSets.

Figure 2. illustrates the mapping of uuCloud

objects to Microsoft Azure and Docker Swarm (Naik,

2016). uuCloud constitutes a middleware layer

between the Microsoft Azure cloud platform and the

Docker cluster orchestration platform Docker Swarm.

uuCloudInstance represents a private or public cloud

instance. There is one to one mapping between Azure

Regions, Resource Groups and virtual Machines, and

the corresponding uuCloud objects: uuRegions,

uuResourceGroups and uuHosts, and Swarm Clusters

and Swarm Nodes that map into uuResourceGroups

and uuHosts.

Figure 2: Mapping between uuCloud, Microsoft Azure and Docker Swarm.

Microservices Management with the Unicorn Cloud Framework

829

Figure 3: Mapping of applications (uuApps) to microservices.

The function of the Unicorn Cloud Control Center

(uuC3) is to automate the management and operation

of container-based microservices via a REST API.

uuCloud API includes deployment commands

(deploy, redeploy, undeploy), commands to increase

or decrease the number nodes in a nodeset (scale),

commands that control sharing of applications (share,

unshare) and commands that are used to monitor the

status of nodes and applications (getStatus,

getAttributes, etc.).

uuApp application is composed of sub-

applications (uuSubApps) – independent units of

functionality that implement a particular set of related

business functions (e.g. administration of online

learning courses). As illustrated in Figure 3, each

uuSubApp is implemented as a containerized

microservice (uuAppServer) and is associated with its

own structured or binary storage (uuObjectStore). We

made an architectural decision to implement each

sub-application as a single containerized

microservice to ensure fast access to persistent data

and to maintain security and consistency of the

underlying data sources. In general, each

microservice has its own life-cycle that includes

requirements specifications, design, development,

testing, deployment and monitoring phases. To allow

finer control over computing resource individual

application modules (Separated Performing Parts)

can be directly addressed. This allows selected

modules of the uuSubApp to be deployed on separate

computing nodes improving application scalability

and reducing resource utilization.

Before a microservice can be deployed to a

particular node (or NodeSet) the developer creates a

Node Image from a uuSubApp and a Runtime Stack

that contains all the related archives and components

needed to run the application (i.e. system tools,

system libraries, etc.). The resulting Node Image

constitutes a unit of deployment, its metadata is

recorded in the Operation Registry and the Node

Image is published to the private Docker Image

Registry. During deployment, the uuC3 Deploy

command reads the Deployment Descriptor (JSON

file that contains deployment attributes) and searches

the Operation Registry for an existing application

deployment object.

At runtime, client applications send requests to a

a router via a gateway (uuGateway) that passes each

request to a load balancer. The load balancer selects a

Node from a NodeSet of functionally identical nodes,

optimizing the use of the hardware infrastructure and

providing a failover capability (i.e. if a particular

Node is not responsive the request is re-directed to an

alternative Node).

ICEIS 2020 - 22nd International Conference on Enterprise Information Systems

830

4 CONCLUSIONS

While the use of microservices and container-based

virtualization brings many benefits, the highly

distributed nature of the resulting applications and the

short software release cycles that characterize the

DevOps approach present many challenges to

organizations involved in the development of cloud-

based applications. The management of large-scale

container-based microservices environments requires

automation to ensure fast and predictable application

deployment, auto-scaling and control of resource

usage. Furthermore, there is a need to support the

ever-expanding range of various types of mobile and

IoT devices using cross-platform client-side

application components. Suitable application

development frameworks and associated methods

and tools are an essential pre-requisite for achieving

successful project results on a repeatable basis.

The uuCloud framework described in this paper is

an integral part of the Unicorn Application

Framework that was developed specifically to

address the requirements of modern mobile cloud-

based applications. The UAF is currently used for the

development of large-scale enterprise applications at

Unicorn. An important difference between UAF and

other similar frameworks (e.g. Kubernetes) is that

UAF is not limited to the management of container-

based environments (container orchestration) but it is

a suite of closely integrated frameworks that address

a comprehensive range of requirements of modern

mobile cloud-based applications over the entire

systems development lifecycle. An important benefit

of the uuCloud framework is its ability to manage

complex multi-tenant environments deployed across

multiple (hybrid) cloud platforms, hiding the

heterogeneity of the underlying cloud infrastructures.

uuCloud manages cloud metadata, supports

automatic application deployment and autonomic

scaling of applications. uuCloud multi-tenancy

derives benefits from the economies of scale. With a

large number of tenants deployed on a public cloud

infrastructure such as AWS or Microsoft Azure,

resources can be purchased in bulk and then allocated

in smaller packages to individual tenants who benefit

from the reduced overall cost. Furthermore, uuCloud

includes a detail lifecycle methodology that captures

the knowledge and expertise gained across numerous

projects and describes the entire design and

development process starting with the identification

of business use cases and ending with the

implementation of well-engineered, component-

based software applications.

Developing cloud frameworks such as uuCloud

has its challenges, in particular the need to keep up

with the rapid development and ever-increasing range

of cloud services available from public cloud

providers. Incorporating new types of services, for

example AWS Lambda serverless compute services,

and AWS Machine Learning services, requires

continuous evolution of the framework and

constitutes the focus of our current efforts.

REFERENCES

2017. Microsoft Azure: Cloud Computing Platform &

Services [Online]. Available: https://azure.microsoft.

com/en-au/ [Accessed 22 August 2017 2017].

Agmon Ben-Yehuda, O., Ben-Yehuda, M., Schuster, A. &

Tsafrir, D. 2014. The rise of RaaS: the resource-as-a-

service cloud. Communications of the ACM, 57, 76-84.

Amazon.Com. 2017. http://aws.amazon.com/ [Online].

Available: http://aws.amazon.com/ [Accessed 7 July,

2017 2017].

Balalaie, A., Heydarnoori, A. & Jamshidi, P. 2016.

Microservices architecture enables DevOps: migration

to a cloud-native architecture. IEEE Software, 33, 42-

52.

Baldini, I., Castro, P., Chang, K., Cheng, P., Fink, S.,

Ishakian, V., Mitchell, N., Muthusamy, V., Rabbah, R.

& Slominski, A. 2017. Serverless computing: Current

trends and open problems. Research Advances in Cloud

Computing. Springer.

Beránek, M., Feuerlicht, G. & Kovář, V. Developing

enterprise applications for cloud: the unicorn

application framework. International Conference on

Grid, Cloud and Cluster Computing, GCC, 2017.

Beranek, M., Kovar, V. & Feuerlicht, G. Framework for

Management of Multi-tenant Cloud Environments.

2018 Seattle, USA. Springer International Publishing,

309-322.

Beranek, M., Stastny, M., Kovar, V. & Feuerlicht, G.

Architecting Enterprise Applications for the Cloud: The

Unicorn Universe Cloud Framework. International

Conference on Service-Oriented Computing, 2017.

Springer, 258-269.

Brewer, E. A. Kubernetes and the path to cloud native.

Proceedings of the Sixth ACM Symposium on Cloud

Computing, 2015. ACM, 167-167.

Burns, B., Grant, B., Oppenheimer, D., Brewer, E. &

Wilkes, J. 2016a. Borg, omega, and kubernetes. Queue,

14, 10.

Burns, B., Grant, B., Oppenheimer, D., Brewer, E. &

Wilkes, J. 2016b. Borg, omega, and kubernetes. Queue,

14, 70-93.

Cloudfoundry. 2017. Cloud Application Platform - Devops

Platform, Cloud Foundry [Online]. @cloudfoundry.

Available: https://www.cloudfoundry.org/ [Accessed

28 September 2017].

Microservices Management with the Unicorn Cloud Framework

831

Docker. 2015. What is Docker [Online]. @docker.

Available: https://www.docker.com/what-docker

[Accessed 21 August 2017.

Feller, E., Rilling, L. & Morin, C. Snooze: A scalable and

autonomic virtual machine management framework for

private clouds. Proceedings of the 2012 12th

IEEE/ACM International Symposium on Cluster, Cloud

and Grid Computing (ccgrid 2012), 2012. IEEE

Computer Society, 482-489.

Kumar, R., Gupta, N., Charu, S., Jain, K. & Jangir, S. K.

2014. Open source solution for cloud computing

platform using OpenStack. International Journal of

Computer Science and Mobile Computing, 3, 89-98.

Mahmood, Z. & Saeed, S. 2013. Software engineering

frameworks for the cloud computing paradigm,

Springer.

Mongodb. 2020. MongoDb [Online]. Available: https://

www.mongodb.com/ [Accessed 20 January 2020].

Naik, N. Building a virtual system of systems using Docker

Swarm in multiple clouds. 2016 IEEE International

Symposium on Systems Engineering (ISSE), 2016.

IEEE, 1-3.

Openshift 2020. OpenShift Container Platform by Red Hat,

Built on Kubernetes.

Rimal, B. P., Jukan, A., Katsaros, D. & Goeleven, Y. 2011.

Architectural requirements for cloud computing

systems: an enterprise cloud approach. Journal of Grid

Computing, 9, 3-26.

Schwarzkopf, M., Konwinski, A., Abd-El-Malek, M. &

Wilkes, J. Omega: flexible, scalable schedulers for

large compute clusters. Proceedings of the 8th ACM

European Conference on Computer Systems, 2013.

ACM, 351-364.

Thönes, J. 2015. Microservices. IEEE Software, 32, 116-

116.

Truyen, E., Van Landuyt, D., Lagaisse, B. & Joosen, W.

Performance overhead of container orchestration

frameworks for management of multi-tenant database

deployments. Proceedings of the 34th ACM/SIGAPP

Symposium on Applied Computing, 2019. ACM, 156-

159.

ICEIS 2020 - 22nd International Conference on Enterprise Information Systems

832