Graph Databases Comparison: AllegroGraph, ArangoDB,

InfiniteGraph, Neo4J, and OrientDB

Diogo Fernandes

and Jorge Bernardino

1,2

Polytechnic of Coimbra - ISEC, Rua Pedro Nunes, Coimbra, Portugal

CISUC – Centre for Informatics and Systems of the University of Coimbra, Portugal

Keywords: Graph databases, NoSQL databases, AllegroGraph, ArangoDB, InfiniteGraph, Neo4J, OrientDB.

Abstract: Graph databases are a very powerful solution for storing and searching for data designed for data rich in

relationships, such as Facebook and Twitter. With data multiplication and data type diversity there has been

a need to create new storage and analysis platforms that structure irregular data with a flexible schema,

maintaining a high level of performance and ensuring data scalability effectively, which is a problem that

relational databases cannot handle. In this paper, we analyse the most popular graph databases:

AllegroGraph, ArangoDB, InfiniteGraph, Neo4J and OrientDB. We study the most important features for a

complete and effective application, such as flexible schema, query language, sharding and scalability.

1 INTRODUCTION

Nowadays, data volume is growing exponentially in

social networks, such as Facebook and Twitter,

which store and process massive amounts of data

daily, where this data reaches petabytes of storage.

Current relational databases are predominant in the

market, but their adaptability is poor in processing

connected data. Actual platforms must deal with

huge amounts of data and related information. For

this reason, a special type of databases designated by

graph databases have appeared.

A graph database is a non-relational database

that provides an effective and efficient solution for

the storage of information in current scenarios,

where data is increasingly interconnected. The

storage mechanisms of graph databases are

optimized for graphing, for the way they store

adjacent records linked by direct references. In this

adjacency list, each vertex maintains references to

their adjacent vertices, forming an index species for

the vertices on neighbourhood. This property is

known as index-free adjacency (Robinson et al.,

2015). Interest in graph models has been increasing

in recent years due to their applications in areas like

Semantic Web and Social Network Analysis

(Dietrich et al., 2008). This type of database is easy

to understand because its concept is based on the

theory of graphs. This theory is basically based on

graphs, which are mathematical structures used to

model relationships between objects. In this context,

a graph is basically a structure, which is represented

by nodes, or also called vertices (the entities), by

edges (the relations) that are the lines that connect

the various nodes and by properties (attributes).

Therefore, the graph databases can simply be

described as a way of representing and storing data,

using their structures: nodes, edges, and properties.

For that reason, graph databases are optimized for

storing and querying graphs.

The problem with graph databases is that they

are not particularly efficient in all desired operations

such as in the representation of data that are derived

from relational models. Therefore, they do not

replace relational databases but are in fact an

efficient solution when dealing with huge volumes

of data that contain many related data.

The focus of this paper is the study of the main

characteristics, advantages and use cases of graph

databases. We study five of the most popular graph

databases: AllegroGraph, ArangoDB, InfiniteGraph,

Neo4J and OrientDB.

The rest of this paper is structured as follows.

Section 2 describes the advantages and uses of graph

databases. Sections 3 to 7 describe AllegroGraph,

ArangoDB, Infinite Graph, Neo4J and OrientDB,

respectively. Section 8 presents the comparison of

the five graph databases. Finally, Section 9 presents

the main conclusions and future work.

Fernandes, D. and Bernardino, J.

Graph Databases Comparison: AllegroGraph, ArangoDB, InﬁniteGraph, Neo4J, and OrientDB.

DOI: 10.5220/0006910203730380

In Proceedings of the 7th International Conference on Data Science, Technology and Applications (DATA 2018), pages 373-380

ISBN: 978-989-758-318-6

373

2 MAIN ADVANTAGES AND

USES OF GRAPH DATABASES

Graph databases are getting popular as data

increases in scalability and relationships become the

first choice over the relational databases. Ironically,

legacy relational database management systems

(RDBMS) are poor at handling relationships

between data points. Their tabular data models and

rigid schemas make it difficult to add new or

different kinds of connections. Graphs are the future.

Not only do graph databases effectively store the

relationships between data points, but they are also

flexible in adding new kinds of relationships or

adapting a data model to new business requirements

(Webber et al, 2015). Graph databases did not see a

greater advantage over relational databases until

recent years, when frequent schema changes,

managing enormous volume of data, real-time query

response time and more intelligent data activation

requirements, make people realize the advantages of

the graph model.

This technology is disrupting many areas, such

as supply chain management, e-commerce

recommendations, security, fraud detection and

many other areas in advanced data analytics.

The main advantages of a graph database are the

following (Guia et al., 2017) and (Neo4J, 2018):

 Better optimization on gathering information

than relational databases;

 ACID rules support;

 Support of the data storage in the order of

petabytes (10

);

 Allow for new types of data;

 Easily expandable due to its structural model -

graph model;

 Suitable for complex and irregular data,

normally involved in real world;

 Excellent for data mining operations;

 A high performance while querying deep data

when compared to relational databases;

 There is no need to declare the data type for the

nodes or edges, unlike relational model;

 They are very agile in development, since they

can be easily adapted over time;

 Combination of multiple dimensions to manage

big data, including time series, geo-dimensions,

and hierarchies on different dimensions.

In short, the three key advantages are:

performance, flexibility, and agility.

Performance, because with traditional databases

relationship queries will come to a grinding halt as

the number and depth of relationships increase. In

contrast, graph database performance stays constant

even as our data grows year after year.

Flexibility, because IT and data architect teams

move at the speed of business because the structure

and schema of a graph model adjusts itself as

applications and industries change. Rather than

exhaustively modeling a domain ahead of time, data

teams can add to the existing graph structure without

endangering current functionality.

Finally, agility because developing with graph

databases aligns perfectly with today’s agile, test-

driven development practices, allowing graph

database to evolve in step with the rest of the

application and any changing business requirements.

However, as any new technology replacing old

technology, there are still obstacles in adopting

graph databases. One is that there are fewer qualified

developers in the job market than the SQL

developers. Another is the non-standardization of the

graph database query language (Wu, 2017).

In the next sections, we analyze five of the most

popular graph databases.

3 AllegroGraph

AllegroGraph is a modern, consistent and persistent

Resource Description Framework (RDF) graph

database with high performance that is currently in

use in open source, commercial, and US Department

of Defense projects. AllegroGraph is characterized

by the efficient use of memory by combining disk

storage, making it possible to scale up to one billion

nodes, always maintaining top performance.

Basically, it provides services including vision

building, rapid prototyping and proof-of-concept

development, complete enterprise technology

solution stack, and best practices to maximize value

from semantic technologies.

AllegroGraph provides an architecture through

the REST protocol, which is an architectural style

that consists of a coordinated set of constraints

applied to components, connectors, and data

elements within a distributed system. Developers

have developed adapters for the various supported

languages, Sesame Java, Sesame Jena, Python, using

proprietary Sesame and Lisp signatures.

AllegroGraph’s competitive advantages are the

following (Ranking of Graph DBMS, 2018):

 Suited to support adhoc queries through

SPARQL, Prolog and languages like

JavaScript;

 Sorted quintuple indices that will index every

primary and non-primary field so users never

DATA 2018 - 7th International Conference on Data Science, Technology and Applications

374

have to worry about whether a certain field is

indexed or not;

 It is possible to mix Geospatial, Temporal,

Social Network Analytics, and Reasoning, all

in the same query (SPARQL or Prolog);

 Triple Level Security with Security Filters;

 Gruff - Graph Visualization, Generate

SPARQL and Prolog queries visually;

 Full and Fast Recoverability;

 Online Backups, Point-in-Time Recovery,

Replication, Warm Standby;

 SOLR and MongoDB Integration;

 All clients based on REST Protocol;

 Cloud-Hosted AllegroGraph - Amazon EC2.

One of AllegroGraph limitations is it focuses on

geo-temporal reasoning and social network analysis.

Figure 1 shows an example of Allegro Graph

interface. In the middle, we have a graph database

and on the left we have subtitles for the nodes and

edges colours.

Figure 1: AllegroGraph interface. (AllegroGraph interface,

no date).

4 ArangoDB

ArangoDB is a multi-model database system

developed by triAGENS GmbH. The data can be

stored as key or value pairs, documents or graphs

and all of this can be accessed by just one query

language (AQL - ArangoDB Query Language).

ArangoDB uses the same core and same query

language for all the data models. If a new product is

being developed and then new ideas are generated,

the model requires necessary changes frequently and

ArangoDB provides the opportunity of keeping the

product updated. Combining different data models in

one query makes data less complicated. This multi-

model makes sense because it has simplified

performance scaling, increased flexibility, fault

tolerance, huge amount of storage memory and

lower cost than other databases. This software has

two versions: Community and Enterprise editions.

The ArangoDB cluster architecture is a master /

master CP (Consistent and Partition) (Mehra, 2017)

model without a single point of failure. With "CP",

we mean that in the presence of a network partition,

ArangoDB prefers to maintain its internal

consistency instead of availability. Clients will have

the same view of the database regardless of the node

to which they connect, and the cluster continues to

fulfil requests even when a machine fails. In this

way, ArangoDB was designed as a distributed multi-

model database (ArangoDB, no date).

The most important advantages and features of

ArangoDB are:

 Management of multiple data models with a

single core and query language (AQL);

 HTTP API to manage database;

 Multi architecture – unique instance, cluster or

mixed services;

 JavaScript Foxx Framework integration;

 Sharding;

 Cloud ready on Azure Cloud;

 Consolidation - minimizes the components,

reducing the complexity of the technology

stack. This means a lower total cost of

ownership, increasing flexibility and

consolidating technical needs;

 Simplified Performance Scaling - can easily

react to growing performance and storage

needs by independently scaling with different

data models. ArangoDB scales both vertically

and horizontally, and if our performance needs

decrease, we can easily scale down the back-

end system to save on hardware and

operational requirements;

 Reduced Operational Complexity - Polyglot

Persistence concept is about choosing the right

data model for the right job. A native multi-

model database allows to have polyglot data

without the complexity, but with data

consistency on a fault tolerant system;

 Strong Data Consistency - a single back-end

manages different data models with support for

ACID transactions. It provides strong

consistency on a single instance and atomic

operations when operating in cluster mode;

 Fault Tolerance – it uses a multi-model

database and a consolidated technology stack.

By design, ArangoDB enables modular

architectures with different data models

running and works for cluster usage as well.

Graph Databases Comparison: AllegroGraph, ArangoDB, InﬁniteGraph, Neo4J, and OrientDB

375

It was created in 2012, but the community is still

small. Also, the software complexity may require an

experienced user on database operations despite all

available documentation.

ArangoDB has a web interface as illustrated in

Figure 2. It shows an example of a graph database

with the node labels on the top right corner. On the

left there is a menu with options like cluster

statistics, graph databases collections, query editor,

user information and application services.

Figure 2: ArangoDB interface. (source: docs.arangodb.

com)

5 InfiniteGraph

InfiniteGraph is a distributed graph database

implemented in Java with core in C++ developed by

Objectivity. InfiniteGraph is unique in their massive

scalability, distributed ingest and processing

capabilities and has cloud storage, among other

features. It only offers physical disk storage.

InfiniteGraph has grown at a tremendous pace, and

can be found in areas such as network management

and telecommunications, healthcare, cyber security,

crime prevention, predictive analytics and fraud

detection, bioinformatics, genomics, scientific

research, finance, social CRM applications

supporting enterprise sales and marketing, and social

networking (InfiniteGraph, no date). The physical

model is object-oriented, so two classes are used to

store the vertices and nodes of a base, BaseVertex

and BaseEdge, respectively.

InfiniteGraph is built on a highly scalable,

distributed database where data and processing are

distributed. A single graph database can be

partitioned and distributed across multiple disks and

multiple machines with the ability to query data

between these frontiers. The same database can

access the graph database locally or over a network.

The lock server handles the locking requirements

through database applications, allowing for

simultaneous reading and writing access to the graph

database. Access to the database is not controlled

when a database instance is created, but rather at the

transaction level. Data servers handle remote

databases and application requests for distributed

grid databases and nodes. Although the REST

interface facilitates interactive access to a database

from a browser, it is not suitable for high-

performance access because it deals with many large

graphs.

Following, InfiniteGraph key features and

advantages are described:

 API/Protocols: Java (core C++);

 Multi-property-graph model;

 Online backups;

 Multi-threading processing;

 Cloud enabled;

 Friendly graph visualization tool;

 Parallel query support;

 Export data models to JSON and GraphML;

 Distributed - scalable graph database solution

that distributes the processing load in the most

efficient way for applications;

 Cost Effective - reduces the total cost of

ownership by reducing the need for data

movement and transformations.

InfiniteGraph also has some limitations. It does

not support sharding and does not have a free

version.

In Figure 3, we have an example of the

InfiniteGraph interface.

Figure 3: InfiniteGraph interface.

6 Neo4J

Neo4J is an open-source graph database implement-

ed in Java. The developers describe Neo4J as a fully

transactional database and a persistent Java engine

where we can store structures in the form of graphs

instead of tables. There are many connections inside

DATA 2018 - 7th International Conference on Data Science, Technology and Applications

376

a big data and Neo4J helps find the intertwining

links between them with the help of its visually

interactive graphs. Neo4J uses natives graph storage

which provides the freedom to manage and store

data in a highly disciplined manner. Neo4J is

considered the most popular and used graph

database worldwide, used in areas such as health,

government, automotive production, military area

and others, being a major reference in this area.

This software was created in 2007 and is divided

into three broad versions: a Community Edition that

is a free version, an Enterprise Edition, where there

is a possibility to test a more complete version for 30

days, and there is still a Government Edition, which

is like an upgrade to Enterprise version. This version

is highly focused for government services. The main

differences between the two major versions of Neo4j

are: an existence of online backups, high cache

memory performance, detailed monitoring of the

system, strong management of database locks, and

greater scalability of the database, among other

benefits of the Enterprise Edition (Neo4J, no date).

Neo4J typical application scenarios are real-time

recommendations, identity and access management,

network and IT operations, fraud detection, anti-

money laundering / tax evasion, knowledge graphs,

graph analytics and algorithms, graph-powered

artificial intelligence, smart homes and IoT.

Currently most databases run through a server,

which can be accessed through a client library. The

Neo4j can run in the embedded mode, as well as in

server mode. Embedded mode must be understood

as an instance that stores all the information on the

disk, and not as an in-memory database, that works

exclusively with the internal memory of the

machine. The embedded Neo4j is ideal for hardware

devices, desktop applications, and for embedded

applications in servers. Running Neo4j in server

mode guarantees a more common way to implement

a database, which is what is most used today. On

each server there is an embedded instance of Neo4j.

Neo4J has many competitive advantages, which

makes this software one of the most used ones in

this area. Following, we describe the major features

of Neo4J:

 Flexible schema;

 Follows property graph data model;

 Scalability and reliability;

 Cypher query language;

 HTTP API to manage the database;

 Index support using Apache Lucence;

 Drivers support like Java, Spring, Scala,

JavaScript;

 Online backups;

 Cloud enabled;

 Exporting of query data to JSON and XLS

format;

 Most active graph community in the world;

 High performance thanks to native graph

storage and processing;

 Easy to learn and to use;

 Easy to load data into the software;

 Whiteboard-friendly data modelling to

simplify the development cycle.

Neo4J doesn’t support sharding and the

Community version, being a free version, it has

some limitations on the number of nodes,

relationships and properties.

In Figure 4 we can see an example of the Neo4J

interface running on a web browser. The sidebar on

the left shows database information, node labels,

relationship types and property keys. In the centre

we have an example of a graph with multiple nodes

(vertices), the relationships (edges) and attributes

names, and on top we have text section where we

can query data in Cypher language.

Figure 4: Neo4J interface.

7 OrientDB

OrientDB is a multi-model open source NoSQL

database management system that supports data

models in documents, graphs, key/value, and

objects.

It was launched commercially in 2011 by

OrienTechnologies and implemented in Java. It is

transactional and supports distributed architecture

with replication. The manipulation of the database

can be done in Java, SQL or with Gremlin. Physical

data storage can be done in memory and on disk.

Like all systems, it uses the free adjacency list to

enable native query processing, but unlike the

Graph Databases Comparison: AllegroGraph, ArangoDB, InﬁniteGraph, Neo4J, and OrientDB

377

others, it uses document database and object-

orientation capabilities to store physical vertices.

OrientDB supports schema less, full and mixed

modes including SQL as one of the languages used

and replication and sharding that can be used in most

complex use cases. OrientDB provides safety in all

confidential data that is present with the use of

authentication, password and data-at-rest encryption.

It has a Community Edition version that is free

(Apache 2 License) but does not support features

such as horizontal scalability, fault tolerance,

clustering, sharding and replication. Enterprise

Edition is an extension and supports features such as

query profiles, distributed cluster configuration,

metric record, live monitoring, a migration tool and

alert configuration. OrientDB use cases are fraud

detection, network/IT operations, graph search,

recommendation engines, master data management,

identity management and forensic analysis.

The OrientDB architecture occurs in distributed

mode. This can be distributed across different

servers and used in diverse ways to achieve

maximum performance, scalability and robustness.

OrientDB uses the Hazelcast Open Source

project for automatic node discovery, storing run-

time cluster configuration, and synchronizing certain

operations between nodes.

Following, we present the key features and

advantages of OrientDB:

 Supports SQL language;

 Web technologies support – HTTP, RESTful

protocol, JSON libraries;

 Distributed – multi-master replication support;

 Cloud ready;

 Database manipulation using Java;

 It can embed documents like any other

document database but also supports

relationships;

 Multi-master plus sharded architecture,

providing horizontal scalability and reliability;

 Fast installation;

 Free version using Apache Licence 2.

OrientDB have some limitations such as not

having an import tool. Sometimes the documentation

is outdated and some users have experienced some

bugs with the graph editor.

OrientDB also provide a web interface. Figure 5

shows an example of a graph database with a sidebar

bar, which shows nodes and edges properties, and

some edition functions like add vertices or clear

canvas on top right corner.

Figure 5: OrientDB interface. (OrientDB, no date).

8 COMPARISON OF FEATURES

After analysing the five graph databases and once

the functionalities are the most important properties

when choosing a graph database, a summary was

made to represent the quality of each feature in each

graph database. Based on (Buerli, 2012) and (Cox,

2017) we select the most important features that a

graph database should have:

 Flexible schema - While relational databases

require new tables or alterations in the existing

ones to add new types of data, in a graph

database we can add new type of vertices and

edges without alterations in the previously

stored data;

 Query language – Relational databases use SQL

query language to manipulate the database, but

graph databases needed a more powerful query

language. There are many different query

languages to manipulate specific graph

databases like Cypher on Neo4J or AQL on

ArangoDB, and we will study to see which one

suits best the needs of querying data;

 Sharding - One of the key elements to being

able to scale NOSQL databases is sharding,

where a large dataset can be broken up and

distributed across a number of (typically

replicated) shards;

 Backups – Graph databases should provide

functions for planning, performing and

restoring a database backup. A full backup

contains all data files and information required

to restore a repository to the state it was in at

the time of the backup;

 Multi-model – A multi-model graph database

provides a database with unstructured data, and

we can visualize relationships with data like in

the form of graphs, key-value pairs, documents

or tables;

DATA 2018 - 7th International Conference on Data Science, Technology and Applications

378

 Multi-architecture - When planning a graph

database solution, there are several structural

decisions to make. These decisions may vary

slightly, depending on which database product

you choose. We will analyse which graph

databases give the best options for the

architecture to be implemented;

 Scalability - As hardware continues to innovate

at a rapid pace (mass storage devices, CPU and

networks), this in turn leads to increased

capacity for existing database software,

enabling growth in the number of transactions

per second. There are two approaches to

scaling a database: Vertical (Scale Up) and

Horizontal (Scale Out). Vertical scaling

involves adding more physical or virtual

resources to the underlying server hosting the

database – more CPU, more memory or more

storage. Horizontal scaling involves adding

more instances/nodes of the database to deal

with increased workload;

 Cloud ready – A graph database implemented

on cloud is a great feature because it solves

scalability problems and provides real-time

management.

Table 1 summarizes our study, where each row

represents the most important features to take into

consideration when choosing a graph database.

In this comparison a five Likert scale from 0 to 4

was used. The grade 4 means that the feature is well

implemented in the software and 1 means that the

feature is not well implemented and should be

improved. 0 (zero) is assigned if the feature is not

supported by the software. The grades are assigned

according to our experience and literature review.

Following, we present the legend for comparison:

Great: 4 points; Good: 3 points; Average/Normal: 2

points; Bad: 1 point; Does not support: 0 points.

Table 1: Graph databases features comparison.

The analysis of AllegroGraph demonstrates that

this software has good features. However,

comparing to the other graph databases, some

features are worst implemented on AllegroGraph. It

also has a architecture supported by REST protocol

and a query language named SPARQL. It is a

language designed for querying RDF graphs but

does not seem to be widely adopted on "SQL of

Graphs". It is used more widely for querying larger

data sets than closed local datasets. One of the main

weaknesses is that the cloud support is still in beta

version and does not provide a flexible schema.

ArangoDB has more functionalities than

AllegroGraph. It provides the user the chance to use

a flexible schema or not; on the other hand it is a

schema-free database and thus allows flexible

storage. All documents in a collection can have the

same or totally different structures. ArangoDB

exploit the similarities in document structures to

save storage space. It will detect identical document

schemas and will only save each unique schema

once. This process is called shaping in ArangoDB.

ArangoDB has a SQL like, powerful and modern

language named AQL that enables really impressive

complicated queries and processes to be pushed to

the backend, which makes this a big strength. It also

has functionalities like sharding, scalability, which is

a multi-model database, and runs in the AzureCloud.

The information available on ArangoDB backups is

a bit confusing, with the community asking many

questions, so it’s a feature that they should improve.

Next, we have InfiniteGraph with functionalities

like flexible schema, cloud ready, a well

implemented backup tool, and it uses Gremlin

language for querying data, a popular language for

people working with graph databases. On the other

hand, this software does not support sharding and

uses only graph data models.

Neo4J is one of the best options when choosing a

graph database and this study shows that this

software has the most important features like

flexible schema, a powerful query language named

Cypher, a tool for planning, performing and

restoring backups, two types of architecture (server

mode and embedded mode), scalability and cloud

prepared. Neo4J does not support sharding, and that

is the main weakness of this software.

Finally, we have OrientDB. This software can

work in a schema-less mode but also supports both

schema-full and schema-hybrid solutions. OrientDB

uses Java, SQL and Gremlin to manipulate data,

supports sharding, can be implement in the cloud

and has great scalability once it uses the distributed

architecture. We could not find clear information on

how the backups work.

After this analysis, we can conclude that Neo4J

and ArangoDB stand out for their functionalities

with 26 points, and is the best graph databases

Graph Databases Comparison: AllegroGraph, ArangoDB, InﬁniteGraph, Neo4J, and OrientDB

379

option nowadays. We can use Neo4J that is

optimized for graph databases and ArangoDB and

OrientDB for databases with different data models

(graphs, documents, key-value). However, we

recommend that Neo4J is a better software with

enhanced implemented features over ArangoDB.

Some websites like db-engines (Ranking of

Graph DBMS, 2018), that study databases ranking

according to their popularity, and Predictive

Analytics Today (Top Graph Databases, 2017), that

study trending graph databases, also have in their

analysis Neo4J, ArangoDB and OrientDB on the top

of their charts with Neo4J as the first choice in both.

This makes this graph database the most used one in

the world.

9 CONCLUSIONS AND FUTURE

WORK

Graph databases provide more performance,

flexibility and agility than non-relational databases.

In this work, we analysed five of the most

commonly used graph databases: AllegroGraph,

ArangoDB, InfiniteGraph, Neo4J and OrientDB.

We conclude that although they all have the

same purpose, which is storing large volumes of data

with many relationships between them, they are very

different in the functionalities they offer. For

example, Neo4J can be implemented in embedded

mode or server mode, and the OrientDB can only be

implemented in distributed mode. These five graph

databases have some distinct advantages and

features, but since the purpose of the software is the

same, there are also many similarities.

When choosing a software to serve specific

needs, we will have to see what each software offers,

and graph databases are no exception. The

functionalities that each one offers is what makes

one better than the other and so an analysis was

made for the features of these graph databases.

We conclude that Neo4J and ArangoDB offer the

best functionalities to implement a graph database

with Neo4J for standing out for its simplicity and

due to its powerful query language named Cypher

despite the need to have previous knowledge of it to

manipulate the database. It also offers two

architecture options, an intuitive interface and a

flexible schema that allows the user to access very

specific information very fast. ArangoDB is also a

powerful tool, has an easy-to-learn and powerful

query language (AQL), and a flexible schema.

Furthermore, since it is a multi-model graph

database, it provides a lot of flexibility and supports

sharding. These are two advantages over Neo4J and

this is why it is also a good choice.

As future work, we intend to analyse the two best

graph databases of this study - Neo4J and

ArangoDB - in a real environment, using a

benchmark to evaluate performance.

REFERENCES

AllegroGraph interface, http://franz.com/ps/newsletter-

archive/32.png

AllegroGraph, https://franz.com/agraph/allegrograph/

ArangoDB, https://www.arangodb.com/

Buerli, M., 2012, “The Current State of Graph Databases”,

Department of Computer Science, Cal Poly San Luis

Obispo, California.

Cox, G., 2017, “Introduction to Graph Databases”,

https://www.compose.com/articles/introduction-to-

graph-databases/

Dietrich, J., Jones, N. and Wright, J., 2008, “Using social

networking and semantic web technology in software

engineering – Use cases, patterns and a case study”,

Massey University Institute of Information Sciences

and Technology, Palmerson North, New Zeland,

January 2008.

Guia, J., Soares, V. and Bernardino, J., 2017, “Graph

Databases: Neo4J Analysys”, 19th International

Conference on Enterprise Information Systems.

InfiniteGraph interface, http://image.slidesharecdn.com/

igoverviewdbtaaug2011-110830134434-phpapp02/95/

webinar-an-introduction-to-infinitegraph-and-connect

ing-the-dots-in-big-data-17-728.jpg?cb=1314711996

InfiniteGraph, http://www.objectivity.com/products/infini

tegraph/

Mehra, A., 2017, “Understanding the CAP Theorem”,

https://dzone.com/articles/understanding-the-cap-

theorem

Neo4J, “Why Graph Databases?”, https://neo4j.com/why-

graph-databases/, (Accessed: 2 of March of 2018)

Neo4J, www.neo4j.com

OrientDB, https://orientdb.com/

Ranking of Graph DBMS, 2018, db-engines,

https://db-engines.com/en/ranking/graph+dbms

Robinson, I., Webber, J. and Eifrem, E., 2015, Graph

Databases, 2

Edition, O‟Reilly Media Inc.,

California

Top Graph Databases, 2017, Predictive Analytics Today,

https://www.predictiveanalyticstoday.com/top-graph-

databases/

Wu, M., 2017, “What Are the Major Advantages of Using

a Graph Database?”, https://dzone.com/articles/what-

are-the-pros-and-cons-of-using-a-graph-databa

DATA 2018 - 7th International Conference on Data Science, Technology and Applications

380