Data Warehousing in the Cloud: Amazon Redshift vs Microsoft

Azure SQL

Pedro Joel Ferreira

, Ana Almeida

and Jorge Bernardino

Instituto Superior de Engenharia do Porto, Rua Bernardino de Almeida, Porto, Portugal

Instituto Superior de Engenharia de Coimbra, Rua Pedro Nunes, Coimbra, Portugal

Keywords: Cloud Computing, Data Warehousing, Cloud Data Warehousing.

Abstract: A data warehouse enables the analysis of large amounts of information that typically comes from the

organization's transactional systems (OLTP). However, today's data warehouse systems do not have the

capacity to handle the massive amount of data that is currently produced, then comes the concept of cloud

computing. Cloud computing is a model that enables ubiquitous and on-demand access to a set of shared or

non-shared computing resources (such as networks, servers, or storage) that can be quickly provisioned or

released only with a simple request and without human intervention. In this model, the features are almost

unlimited and in working together they bring a very high computing power that can and should be used for

the most varied purposes. From the combination of both these concepts, emerges the cloud data warehouse.

It advances the way traditional data warehouse systems are defined by allowing their sources to be located

anywhere as long as it is accessible through the Internet, also taking advantage of the great computational

power of an infrastructure in the cloud. In this paper, we study two of the most popular cloud data

warehousing market solutions: Amazon Redshift and Microsoft Azure SQL Data Warehouse.

1 INTRODUCTION

Data warehouses (DW) are defined as customized

data storage that aggregate data from multiple

sources and store it in a common location to be able

to run reports and queries over it. This concept arose

from the requisite to integrate corporate data across

multiple application servers that an organization

might have, so that it would be possible to make data

accessible to all users who need to consume

information and make decisions based on it. Many

companies use data warehouses to compile regular

financial reports or business metric analyses.

Integration of a Cloud Data Warehouse solution

demands a very well defined strategy that would

involve Cloud Computing capabilities. The success

of the implementation depends on the existence of a

service-oriented strategy at the organization level,

which would provide the necessary infrastructure for

the Cloud implementation. Without SOA and BPM,

integration of a Data Warehousing solution based on

Cloud Computing serves no financial purpose,

involving high costs for the present systems

reengineering. Also, in order to be successful, Cloud

strategy has to be led according to the business

strategy of the organization.

One of the biggest challenges with data

warehousing in the Cloud is how the data is

transferred up into it. Pumping gigabytes, terabytes,

or even petabytes of data up into Cloud over the

public Internet can not only come with security

concerns, but also performance challenges.

When selecting a Data Warehousing solution, we

have to take into account the newest trends on the

DW and Cloud Computing market, the present and

future needs and the opportunity of integration.

Therefore, in order to be successful, the selection of

a Cloud DW solution has to be achieved objectively

based on good criteria that have been analyzed and

weighted according to the present and future needs

of the organization. Cloud Data Warehousing is a

potential cost saver for big companies, and removing

a cost barrier that have held data warehousing back

from small and mid-sized businesses.

Cloud Data Warehouse must be designed to take

away the undifferentiated heavy lifting of running

infrastructure at heavy scale, allowing the customers

to focus on their core competencies – its business.

The growing interest in cloud-based data

warehousing is driven by the high return on

investment. Nonetheless, the adoption of cloud

computing for data warehousing faces security

Ferreira P., Almeida A. and Bernardino J.

Data Warehousing in the Cloud: Amazon Redshift vs Microsoft Azure SQL.

DOI: 10.5220/0006587103180325

In Proceedings of the 9th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (KDIR 2017), pages 318-325

ISBN: 978-989-758-271-4

challenges given the proprietary nature of the

enclosed data. Moving to the cloud is a hard

decision not only because the data owners like to

have their data near to them (on-premises) but also

due to security and data confidentiality issues,

because of this last two the decision makers tend to

delay the decision of moving to the cloud.

There are many benefits to implement a DW in

the cloud, such as cost, efficiency, elasticity,

flexibility in the choice of provider, more

competitiveness and less time involved in

installation and maintenance. It is assumed that

cloud system efficiency is even more efficient with

the use of parallel computing and the cloud is

motivation for small and medium enterprises by the

possibility of expansion at an affordable cost.

In this paper we analyze the characteristics of

two of most popular cloud data warehousing

platforms: Amazon Redshift and Microsoft Azure

SQL Data Warehouse. The remainder of this paper

is structured as follows. Section 2 describes the

related work. Section 3 presents the cloud

computing and Section 4 presents a summary of

cloud data warehousing area. Section 5 presents two

cloud data warehousing market solutions: Amazon

Redshift and Microsoft Azure SQL Data Warehouse.

Section 6 presents a brief comparison of the cloud

data warehouse solutions. Finally, conclusions and

future work are summarized in Section 7.

2 RELATED WORK

Multiple research works have been done to compare

and evaluate existing Big Data platforms. However,

most of the research focus only on a specific

capability, technology or purpose.

Almeida and Bernardino (2015a; 2015b) focus

on the capability of mining data, and in a mix of

technical parameters and features that are suitable

for Small and Medium Enterprise environments. On

the other hand, Morshed et al. (2016) focused their

work on platforms addressing distributed real-time

data analytics, and concluded that the platforms

present on their research do not cover all the features

that are required for distributed computation in real-

time.

Kaur et al. (2012) describe some solutions from

multiple vendors in the cloud to support a Data

Warehouse. They state that each service provider

implements different levels of use, some provide

ETL mechanisms or Business Intelligence (BI)

solutions while others provide the storage

infrastructure and the client then chooses the

supplier for their needs and the services required. In

the paper it is considered that the market is still

immature and services vary in price and

performance.

Hemlata Verna (2013) is concerned about how to

manage the data through recycle, reuse, reduce and

recover information in cloud environment. Popeangã

(2014) concentrate her work on studying what archi-

tecture is suited for a data warehouse in the cloud.

Mathur et al., (2011), focused their work on

distributed databases in the cloud. They propose

IaaS cloud servers that can be used to store these

databases at low initial cost.

Therefore, there are few related works which

evaluate solutions based on specific capabilities,

technology or purpose. Our work will be of a

broader scope in functionalities and applications in

order to be used by SMEs.

3 CLOUD COMPUTING

There is no definitive definition of cloud computing,

but Kimball and Ross (2013) propose that “Cloud

computing is a self-provisioning and on-demand de-

livery of computing resources and applications over

the Internet with a pay-as-you-go pricing model”.

Cloud, by definition, is a self-service system that

allows end users to setup applications and services

in a cloud computing environment without the

intervention of an IT service provider (Armbrust et

al., 2010). Cloud computing addresses large amounts

of computing power by aggregating computing

resources and offering a single view of the system.

Cloud brought new features and possibilities to

improve its user’s life but also originated a new

market segment for IT with new business

opportunities. Many organizations build their

business around the cloud not only to use its services

but also to offer business solutions in this

environment. Like any other service, we can identify

at least two actors that are directly connected with it:

• User or consumer: is the one who uses the

functionalities or resources provided by the

cloud. It is the entity or organization that

uses the cloud computing services, whether

they are related with software, platform or

infrastructure;

• Provider: is the entity or organization

responsible for making cloud services

available to consumers. The provider is also

responsible for managing the necessary

infrastructure which supports its services.

There are different cloud implementation models

and they are the different means by which

consumers access cloud services. These models are

different and are characterized by the target audience

they serve. If this assumption is taken into account,

there are only two models: private clouds and public

clouds.

Private clouds are typically used in the

distribution of services in an internal organization

environment. The infrastructure that supports the

private cloud can be delegated to a service provided

and stays private at all, it has not to be acquired and

managed by the private cloud users themselves.

When managed internally by users themselves,

they have full control over the processes, data or

applications used in the cloud. However, they lose

some of the general principals or benefits of cloud

computing, such as access to infrastructure at

reduced prices, elasticity, resources availability or

rapid deployment times.

Unlike private clouds, public clouds are used for

the distribution of services to the general public,

typically via the Internet. The service provider is

responsible for the entire support infrastructure. This

infrastructure is fully shared by the various users of

this cloud. Thanks to this sharing and optimization

of resource management processes, providers are

able to maximize their use and enable them to be

supplied at reduced prices to consumers.

4 CLOUD DATA WAREHOUSING

Nowadays, companies have a greater collection of

data than ever before. This includes a huge variety

of sources, including cloud-based applications or

even company datamarts. In order to make good

decisions, get insights and achieve a competitive

advantage, companies need to have their data

properly analyzed, on time.

The conventional data warehouse architecture is

widespread in a large number of companies working

with massive and diverse data sets but is a very

closed and complex model to respond with the

agility that companies currently need (Tereso and

Bernardino, 2011). People who make analysis need

to wait a number of hours or even days for data to

flow into the data warehouse before it becomes

available for analysis. In most cases, the storage and

compute resources required to process that data are

insufficient (or the same) and this leads to hanging

or crashing systems (Goutas et al., 2016).

One of the major concerns on moving to the

cloud is the time to “live” with both on-premises and

cloud data warehouse system because it is not a

good idea to move at once the whole data ware-

house. To ease this concern, a data virtualization

solution can be used to help out the migration and

coexistence of the both data warehouse systems

while migration to cloud is ongoing.

Cloud data warehousing was born from the

convergence of three trends – huge changes in data

sources, volume and complexity; the need for data

access and analytics; and better technology that

increased the efficiency of data access, analytics and

storage. Traditional data warehouse systems were

not designed to handle the volume, variety and

complexity of today’s data (Almeida et al., 2008).

A data warehouse in the cloud is a database which

information is consumed over the Internet, a typical

database as a service (DBaaS). Cloud data ware-

housing is a cost-effective way for organizations to

use and take advantage of high technology without

high upfront costs to purchase, install and configure

the required hardware, software and infrastructure

(Talia, 2013).

In the next section we will analyse two of the

most popular cloud data warehousing market

solutions.

5 CLOUD DATA WAREHOUSING

MARKET SOLUTIONS

In this section, the architecture of two cloud data

warehousing solutions is described.

The following sections describe the characteris-

tics of most popular platforms: Amazon Redshift

and Microsoft Azure SQL Data Warehouse.

5.1 Amazon Redshift

Gartner reports Amazon Web Services (AWS) is

often considered the leading cloud data warehouse

platform-as-a-service provider (Gartner, 2016a).

Recognized by Gartner as a leader, Amazon

Redshift is a fast, fully managed, petabyte-scale data

warehouse that makes simple and cost-effective to

analyze all our data using existing business

intelligence tools.

Amazon Redshift engine is a SQL-compliant,

MPP, query processing and database management

system designed to support analytics workload. The

storage and compute is distributed across on or more

compute nodes (Gupta et al., 2015).

The core infrastructure of Amazon Redshift data

warehouse is a cluster and it is composed by:

• A leader node;

• One or more compute nodes.

The leader node accepts connections from the

client applications and dispatch the work to the

compute node: it parses and develops execution

plans to carry out database operations and based on

the execution plan it compiles code, distributes the

compiled code to the compute nodes and assigns a

portion of the data to each node (see Figure 1).

The leader node distributes SQL statements to

the compute nodes only when a query references

tables that are stored on the compute nodes,

otherwise they run exclusively on the leader node

(“Data Warehouse System Architecture - Amazon

Redshift,” n.d.).

Figure 1: Amazon Redshift system architecture (source:

aws.amazon.com/redshift).

The compute nodes execute the compiled code

sent by the leader and send the results back for final

aggregation. Each compute node has its own

dedicated CPU, memory and storage – it is easy to

scale the cluster by upgrading the compute nodes or

adding new ones. The minimum storage for each

compute is 160GB and scale up to 16TB to support a

petabyte of data or more. The compute node is

partitioned into slices and each of it is allocated a

portion of the node’s memory and disk space, where

it processes a portion of the workload assigned to the

node – the leader node manages the distribution of

data of the workload to the slices and then they work

in parallel to complete the operation.

The cluster contains one or more databases.

Amazon Redshift is a relational database

management system and provides the same

functionality as a typical RDBMS including all

related with OLTP but it is optimized for high-speed

performance analysis and reporting of very large

datasets (“Data Warehouse System Architecture -

Amazon Redshift,” n.d.).

The database engine is based on PostgreSQL.

Another interesting characteristic of Amazon

Redshift is that it is a columnar database, which

means that each record is not saved as a unique

block of data but it is stored in independent columns.

The query performance can greatly be improved by

selecting a limited subset of columns rather than the

full record.

The data warehouse functionality is comparable

to the high level databases. The ease of use and

scalability of Redshift is definitely a huge advantage

of this solution.

5.2 Microsoft Azure SQL Data

Warehouse

Microsoft Azure SQL data warehouse is a cloud-

based, scale-out database capable of processing

massive volumes of data, both relational and non-

relational. It is a massively processing (MPP)

distributed database system. “It provides SaaS, PaaS

and IaaS services and supports many different

programming languages, tools and frameworks,

including non-Microsoft software (“SQL Data

Warehouse | Microsoft Azure,” n.d.).

SQL Data Warehouse is based on the SQL

Server relational database engine and integrates with

the tool that its users may be familiar with. This

includes (“SQL Data Warehouse | Microsoft Azure,”

n.d.):

• Analysis Services;

• Integration Services;

• Reporting Services;

• Cloud-based tools.

The Microsoft Azure SQL Data Warehouse is

composed by a Control Node, Compute nodes and

Storage. It also has a service called Data Movement

Service that is responsible for the data movement

between the nodes (“SQL Data Warehouse |

Microsoft Azure,” n.d.).

Like the Leader Node of Amazon Redshift, the

Azure Control Node manages and optimizes queries

and is responsible for the coordination of all the data

movement and computation required to run parallel

queries (see Figure 2). When a request is made to

SQL Data Warehouse, the control node transforms it

into separate queries that run on each compute node

in parallel.

The compute nodes are SQL databases that store

the data and process queries. When data is added, it

is distributed to the compute nodes and when the

data is requested these nodes are the workers that

run queries in parallel. After processing, they pass

the results back to the control node so it can

aggregate the results and return the final result to the

user.

All the data stored in Azure SQL Data

Warehouse is stored in Azure Blob Storage – it is a

service that stores unstructured data in the cloud as

objects/blobs. Blob Storage can store any type of

text or binary data, such as a document, media file or

application installer. When compute nodes interact

with data, they write and read directly to and from

blob storage. Compute and Storage are independent.

Figure 2: Microsoft Azure SQL Data Warehouse system

architecture (source: “SQL Data Warehouse | Microsoft

Azure,” n.d.).

As previously referred, Data Movement Service

(DMS) is responsible for all the data movements

between the nodes. It gives the compute nodes

access to data they need for joins and aggregations.

It is not an Azure service but a Windows service that

runs alongside SQL Database on all the nodes and it

is only visible on query plans because they include

some DMS operations since data movement is

necessary to run a query in parallel.

Azure is an enterprise-level SQL data warehouse

that extends the SQL Server family of products and

services into the cloud. Azure can also scale storage

and computing so the customers only have to pay for

what they require.

6 COMPARING CLOUD DATA

WAREHOUSES

Data warehouses in the cloud are gaining popularity

because cloud vendors offer DW services at lower

costs. While Amazon Redshift is the number one in

the market, according to Gartner (2016), Azure

offers a competitive platform to be considered.

Redshift and Azure SQL Data Warehouse both

support petabyte scale systems. Both of them have

leader or control nodes and compute nodes. The

biggest difference between Azure SQL Data

Warehouse and Redshift is the decoupling of storage

and compute resources.

About scalability, in Redshift, when the cluster is

modified, the changes are immediately applied.

While the new clusters are being provisioned, the

current cluster is available in read only mode, so

during this process the data is available for read

operations. After the new clusters are provisioned,

the data is copied.

In Azure SQL Data Warehouse, the scaling of

the clusters can happen in few minutes. The scale

out can be done for compute and storage units

independently. Azure SQL Data Warehouse also

supports pausing a compute operation. There is no

cost applied when the compute nodes are in pause

state; only a storage cost is charged.

From Data sources, data can be integrated with

Redshift from Amazon S3 storage. If there is an on-

premises database to be integrated with Redshift, the

data needs to be extracted from the database to a file

and then imported up into S3.

Azure SQL Data Warehouse is integrated with

Azure Blob storage. It uses a similar approach as

Redshift to import the data from SQL Server. The

SQL Server data is exported to a text file and then

copied across to Azure Blob storage.

Comparing the adoption of public clouds, in

particular AWS and Azure, we can see that AWS is

the number one adopted cloud solution for the most

respondent users of the 2017 Rightscale survey, as

shown in Figure 3.

Figure 3: Adoption of public cloud in 2017 (source:

“RightScale 2017 - State of the cloud report,” 2017).

Although AWS continues to lead in public cloud

adoption (57 percent of respondents currently run

applications in AWS), this number has stayed the

same as in 2016.

Figure 4: Adoption of public cloud in 2017 vs 2016

(source: “RightScale 2017 - State of the cloud report,”

2017).

In contrast, over the last year, happened a

significant growth in the percentage of respondents

running applications in Azure and Google, the

second and third public cloud providers (see Figure

4). Overall Azure adoption grew from 20 to 34

percent of respondents, reducing the AWS lead.

Google also increased from 10 to 15 percent.

A comparison of system properties is shown in

Table 1.

RedShift and Azure SQL have a database model

based on relational database management system

(RDBMS) which supports the relational data model.

Amazon Redshift is built around industry-

standard SQL, with added functionality to manage

very large datasets and support high-performance

analysis that data. Although it is based on

PostgreSQL, there are some unsupported features,

data types and functions. Some SQL features are

also implemented differently, for example:

• CREATE TABLE

• ALTER TABLE

• INSERT, UPDATE and DELETE

Amazon Redshift does not support tablespaces,

table partitioning, inheritance, and certain

constraints. The Amazon Redshift implementation

of CREATE TABLE enables users to define the sort

and distribution algorithms for tables to optimize

parallel processing. ALTER COLUMN actions are

not supported. ADD COLUMN supports adding

only one column in each ALTER TABLE statement.

Using INSERT, UPDATE, and DELETE the WITH

is not supported. For the complete list of

unsupported features, data types and functions, our

suggestion is to check on AWS documentation, in

particular the one that concerns about Amazon

Redshift and PostgreSQL (“Amazon Redshift and

PostgreSQL - Amazon Redshift,” n.d.).

Table 1: Comparison Redshift Vs Azure SQL Data

Warehouse.

System

Properties

Amazon

Redshift

Microsoft

Azure SQL

Data

Warehouse

Database

model

Relational

DBMS

Relational

DBMS

Developer

Amazon

(based on

PostgreSQL)

Microsoft

Licence Commercial Commercial

Cloud based Yes Yes

Implementation

language

C C++

XML support No Yes

SQL standard

support

Does not fully

support

Yes

Supported

programming

languages

All languages

supporting

JDBC/ODBC

.Net, Java,

JavaScript

PHP, Python

Ruby

Server-side

scripts

User defined

Python

functions

Transact SQL

Support for

concurrent data

manipulation

Yes Yes

MapReduce

API support

No No

In-memory

support

Yes No

Control over

node

configuration

Yes No

In-Memory OLTP is a technology for optimizing

performance of transaction processing, data

ingestion, data load, and transient data scenarios. In-

Memory support is available on RedShift but not in

Azure SQL Data Warehouse because it is only

available for OLTP workloads in SQL Server since

version 2014 and Azure SQL Database.

In MapReduce support property, both databases

do not offer an API for user-defined Map/Reduce

methods. In case of RedShift, it is possible to

combine MapReduce with RedShift by processing

input data with MapReduce and import results to

RedShift. In case of Azure SQL Data Warehouse,

Polybase unifies data in relational data stores with

non-relational ones, combining data from both

RDBMS and Hadoop so that users don’t need to

understand HDFS or MapReduce.

Redshift and Azure SQL Data Warehouse offer

many similar capabilities, so it is not necessarily a

matter of one provider being better or worse than the

other. It all depends on what our business needs, but

each solution has pros and cons. See Table 2 to find

some pros and cons of Amazon RedShift and

Microsoft Azure SQL Data Warehouse cloud

solutions.

Table 2: Pros and Cons of Redshift and Azure SQL Data

Warehouse.

Pros Cons

Amazon

Redshift

Performance

through use of

local storage

Compute cannot

be scaled

independent of

storage (and vice

versa)

Loading data

from S3 is very

fast

Can't pause

resources

Beyond Petabyte

Queries that

require joins

against multiple

columns can

suffer in

performance

Columnar data

store allows high

performance

queries on large

volumes of data

Familiarity with

PostgreSQL

makes adopting

Redshift easier

Microsoft

Azure SQL

Data

Warehouse

Resources can be

paused during

idle time in

workload

Can only run 32

concurrent

queries

(maximum)

Scale separate

compute and

storage resources

and pay only for

what is used

Not fully

supports T-SQL

Excellent

integration with

Azure Services

7 CONCLUSIONS AND FUTURE

WORK

Data warehouses are defined as customized data

storage that aggregate data from multiple sources

and store it in a common location to be able to run

reports and queries over it. Many companies use

data warehouses to compile regular financial reports

or business metric analyses.

In this paper we analyse cloud data warehousing

area that is the convergence of three trends – huge

changes in data sources, volume and complexity; the

need for data access and analytics; and better

technology that increased the efficiency of data

access, analytics and storage. Traditional data

warehouse systems were not designed to handle the

volume, variety and complexity of today’s data.

The integration of a Cloud DW solution needs a

very well defined strategy that would involve Cloud

Computing capabilities. The success of the

implementation depends on the existence of a

service-oriented strategy at the organization level,

which would provide the necessary infrastructure for

the Cloud implementation.

In this study we conclude that there are several

challenges when deploying data warehouses into the

cloud:

• Importing data for the data warehouse into

the cloud for storage can be a challenge,

because when using the cloud, a customer

is dependent on the Internet connection and

the infrastructure of the cloud provider. It

can be necessary to use a dedicated

communication line to mitigate the

connection problems but it as a cost;

• Getting large amounts of data from cloud

storage to compute nodes provided by the

cloud solution for computing can lead to a

performance issue;

• Loss of control can lead to issues involving

security and trust.

In this work we analyse and evaluate two of the

most popular cloud data warehousing solutions:

Amazon Redshift and Microsoft Azure SQL Data

Warehouse.

Some conclusions concerning the two cloud data

warehouse solutions have been taken:

• Using Redshift to scale our data warehouse

we must increase both the compute and

storage units. With Azure SQL DW,

compute and storage is decoupled so we

can scale them individually. This is a very

different economic model that can save

customers a lot of money as they don’t

have to purchase additional storage when

they just need more compute power, or

vice-versa.

• Azure SQL DW has the ability to pause

compute when not in use so we only pay

for storage, as opposed to Redshift in which

we are billed 24/7 for all the virtual

machines that make up the nodes in our

cluster.

• RedShift is easier to configure than Azure

SQL Data Warehouse and takes less time to

be online and available after its setup.

As future work we intend to analyze these two

platforms with data from a company, and a

recommendation will be given of what is the best

cloud data warehouse solution in the market based

on a set of criteria.

REFERENCES

Almeida, R., Vieira, J. Vieira, M. Madeira, H. and

Bernardino, J. “Efficient Data Distribution for DWS”.

In International Conference on Data Warehousing

and Knowledge Discovery - DaWaK, pages 75–86,

2008.

Almeida, P., and Bernardino, J. “Big Data Open Source

Platforms”. BigData Congress 2015: 268-275

Almeida, P., and Bernardino, J. "A comprehensive

overview of open source big data platforms and

frameworks", International Journal of Big Data

(IJBD), 2(3), 2015, pp. 15-33.

Amazon Redshift and PostgreSQL - Amazon Redshift

[WWW Document], n.d. URL http://docs.aws.

amazon.com/redshift/latest/dg/c_redshift-and-

postgres-sql.html (accessed 9.2.17).

Amazon Redshift vs. Microsoft Azure SQL Data

Warehouse vs. Microsoft Azure SQL Database

Comparison [WWW Document], n.d. URL https://db-

engines.com/en/system/Amazon+Redshift%3BMicros

oft+Azure+SQL+Data+Warehouse%3BMicrosoft+Az

ure+SQL+Database (accessed 9.2.17).

Armbrust, M., Fox, A., Griffith, R., Joseph, A.D., Katz,

R., Konwinski, A., Lee, G., Patterson, D., Rabkin, A.,

Stoica, I., Zaharia, M., 2010. A View of Cloud

Computing. Communications of ACM 53, 50–58.

Combining Hadoop/Elastic Mapreduce with AWS

Redshift Data Warehouse [WWW Document], n.d.

URL http://atbrox.com/2013/02/25/combining-hadoop

elastic-mapreduce-with-aws-redshift-data-warehouse/

(accessed 9.3.17).

Data Warehouse System Architecture - Amazon Redshift

[WWW Document], n.d. URL https://docs.aws.

amazon.com/redshift/latest/dg/c_high _level_system_a

rchitecture.html (accessed 1.1.17).

Database Manag. Solut. Anal. URL https://www.gartner.

com/doc/reprints?id=1-2ZFVZ5B&ct=160225&st=sb

(accessed 1.2.17).

Gartner, 2016. Magic Quadrant for Data Warehouse and

Database Management Solutions for Analytics

[WWW Document]. Magic Quadr. Data Wareh.

Goutas, L., Sutanto, J., Aldarbesti, H., 2016. The Building

Blocks of a Cloud Strategy: Evidence from Three

SaaS Providers. Communications of ACM 59, 90–97.

Gupta, A., Agarwal, D., Tan, D., Kulesza, J., Pathak, R.,

Stefani, S., Srinivasan, V., 2015. Amazon Redshift

and the Case for Simpler Data Warehouses, in:

Proceedings of the 2015 ACM SIGMOD International

Conference on Management of Data, SIGMOD ’15.

ACM, New York, NY, USA, pp. 1917–1923.

Hemlata Verna, 2013. Data-warehousing on Cloud

Computing, in: International Journal of Advanced

Research in Computer Engineering & Technology

(IJARCET) Volume 2, Issue 2, February 2013.

Kaur, H., Agrawal P., and Dhiman, A., "Visualizing

Clouds on Different Stages of DWH - An Introduction

to Data Warehouse as a Service," 2012 Int. Conf. on

Computing Sciences, Phagwara, 2012, pp. 356-359.

Key Concepts & Architecture Snowflake Documentation

[WWW Document], n.d. URL https://docs.snowflake.

net/manuals/user-guide/intro-key-concepts.html

(accessed 2.16.17).

Mathur, A., Mathur, M. & Upadhyay, P., 2011. Cloud

Based Distributed Databases: The Future Ahead. In:

International Journal on Computer Science and

Engineering (IJCSE), 3(6), pp.2477-81.

Miller, J. A., Bowman, C., Harish, V.G., Quinn, S., 2016.

Open Source Big Data Analytics Frameworks Written

in Scala, in: 2016 IEEE International Congress on Big

Data (BigData Congress), pp. 389–393.

Morshed, S. J., Rana, J., Milrad, M., 2016. Open Source

Initiatives and Frameworks Addressing Distributed

Real-Time Data Analytics, in: 2016 IEEE

International Parallel and Distributed Processing

Symposium Workshops (IPDPSW). Presented at the

2016 IEEE Int. Parallel and Distributed Processing

Symposium Workshops (IPDPSW), pp. 1481–1484.

Popeangã, J., 2014. Shared-Nothing Cloud Data

Warehouse Architecture, in: Database Systems

Journal vol. V, no. 4/2014.

RightScale 2017 - State of the cloud report [WWW

Document], 2017. URL: https://assets.rightscale.com/

uploads/pdfs/RightScale-2017-State-of-the-Cloud-

Report.pdf (accessed 8.11.17).

SQL Data Warehouse, Microsoft Azure [WWW Docu-

ment], n.d. URL https://azure.microsoft.com/en-

us/services/sql-data-warehouse/ (accessed 11.13.16).

Talia, D., 2013. Clouds for Scalable Big Data Analytics.

Computer 46, 98–101. doi:10.1109/MC.2013.162

Tereso, M., and Bernardino, J. “Open source business

intelligence tools for SMEs”. Information Systems and

Technologies (CISTI), 6th Iberian Conference on,

IEEE (2011) 1–4.