ISSN 2278-3091

Volume Trends
Eduard Shaikhulov, International Journal of Advanced 14, No.4, July - Science
in Computer Augustand
2025
Engineering, 14(4), July – August 2025, 190 - 193
International Journal of Advanced Trends in Computer Science and Engineering
Available Online at http://www.warse.org/IJATCSE/static/pdf/file/ijatcse021442025.pdf
https://doi.org/10.30534/ijatcse/2025/021442025

Fundamentals of working with Big Data in Databases

Eduard Shaikhulov
1
Bachelor’s Degree, Kazan innovative university named after V.G. Timiryasov, Russia, shaihulove@gmail.com

Received Date: June 20, 2025 Accepted Date: July 26, 2025 Published Date: August 06, 2025

 2. TECHNICAL CHARACTERISTICS OF BIG DATA

ABSTRACT
Big data refers to a collection of information arrays
This article examines the key aspects of working with big data characterized by enormous volume, high velocity of incoming
in databases, including their characteristics, architectures, and data and variety of formats [1]. These three aspects define the
processing technologies. It analyzes modern solutions such as requirements for DBMS, which must be adapted to process
distributed computing (Hadoop, Spark), NoSQL databases such information effectively.
(MongoDB, Cassandra, Redis), and relational DBMS (SQL The volume of data in today's systems is measured in
Server, Oracle, PostgreSQL) that are adapted for handling big terabytes, petabytes, and even exabytes because of the
data. The significance of the optimization methods, such as constantly increasing number of information sources such as
indexing, partitioning, and compression, is emphasized, Internet of Things (IoT) sensors, social media, e-commerce
particularly the role played by security controls, such as platforms, and other online spaces. To be able to handle this
encryption, access control, and monitoring. The emphasis is much, a system has to be highly scalable, either by vertically
on integration of the big data with analytical tools with a scaling-increasing one server's capabilities-or horizontally
scalability and high performance orientation. scaling-distributing the load among numerous servers.
The second critical characteristic is the velocity of data
Key words : Big data, databases, database management
generation. A lot of data in modern information systems
systems (DBMS), distributed computing, NoSQL databases.
should be processed in real time or near-real time; this
becomes crucially important in applications like financial
1. INTRODUCTION
transactions, logistics management, and industrial process
The contemporary world has entered the age of digital monitoring.
transformation, with data volumes increasing at an accelerated Data variety is associated with the fact that information comes
rate. Big data is an enormous, unavoidable essential in different formats: structured, semi-structured, and
component of contemporary information systems defined by unstructured. For instance, data can be represented in tables
its size, rapid speed, and variety in formats. It has extended to (relational databases), JSON or XML files (semi-structured
multiple sectors: including healthcare, finance, scientific data), as well as images, videos, or text (unstructured data).
research, commerce, and production. Other important features of big data are veracity, associated
Traditional database management systems (DBMS), designed with trust and quality of the information. In practice, big data
to handle relatively small volumes of structured information, could be incorrect, incomplete, and outdated because the
face significant limitations when working with big data. There sources feeding a central application vary, therefore including
are many more in relation to related limitations with velocity mechanisms of data cleaning, validation, and renewal is
and volume issues, reliability and availability, and handling required.
multi-data format variability. The present interest covers That involves significant restrictions with regard to the
methodologies and technologies specific to big data architectures of database systems, which will surely affect:
processing: distributed computing, NoSQL systems, and thus, they shall be distributed given a large volume, featuring
hybrid database architectures. The aim of this article is to high levels of fault tolerance for instant data access. Another
explore the fundamental concepts of working with big data in no less important side of the story is supportability in
databases. It examines technical characteristics, architectural scalability relevance to analytics with the integration into
solutions, technological tools, and optimization methods used machine learning technologies; databases thus will be
to ensure the efficient management of large volumes of substantial members of recent big data ecosystems.
information.

190
 Eduard Shaikhulov, International Journal of Advanced Trends in Computer Science and Engineering, 14(4), July – August 2025, 190 - 193

3. TECHNOLOGIES FOR BIG DATA Spark an architecture to work with huge volumes of data in
memory with no latency, granting computational performance
Big data necessitates the utilization of specialized to increase in real time.
technologies that can keep pace with such volume, velocity, NoSQL databases are specifically designed to handle
and variety challenges while processing and managing it [2]. unstructured and semi-structured data, which are often
The main enabling technologies in this respect are distributed unsuitable for traditional relational models [4]. MongoDB,
computing, NoSQL databases, and adapted relational DBMS. Cassandra, and Redis are among the most commonly used
Distributed computing is the backbone of big data processing NoSQL solutions. MongoDB is a document-oriented database
systems nowadays. Different platforms, such as Apache that stores data in JSON-like documents, allowing for efficient
Hadoop and Apache Spark, enable the allocation of handling of flexible structures (figure 3).
computational tasks across several nodes to manage large
amounts of data effectively [3]. Hadoop utilizes the
MapReduce framework, where a task is split into smaller
subtasks that are executed simultaneously and subsequently
merged to produce the final outcome (figure 1).

Figure 3: MongoDB architecture diagram

This architecture makes MongoDB particularly valuable in

scenarios where data structures can be changed dynamically
and horizontally scaled. MongoDB, in the diagram, enables
integration with multiple data stores through Mongo
Connector, which tracks changes in the operation log (OP
LOG) and writes them to target systems such as Elasticsearch,
Figure 1: Hadoop workflow diagram Solr, or other MongoDB instances. This feature enables
MongoDB to be effectively employed in distributed systems to
Equally, Hadoop contains the distributed file system HDFS ensure data synchronization, facilitate full-text search
(Hadoop Distributed File System) for reliable data storage capabilities, and support analytics integration.
characterized by high levels of fault tolerance. It is particularly Cassandra, in contrast, is a column-based database featuring
effective in the long-term storage and analysis of huge high scalability and a decentralized architecture. It provides
amounts of data in a distributed environment. high performance and low latency, making it ideal for handling
Spark, on the other hand, offers in-memory data processing substantial data quantities in worldwide distributed networks
capabilities, significantly accelerating task execution, (figure 4).
especially for real-time data analysis or iterative computations
(figure 2).

Figure 2: Spark data processing diagram

Figure 4: Cassandra architecture diagram
Apache Spark is a high-performance big data processing
platform that works with CSV, Sequence File, Avro, and The architecture of Cassandra is based on a cluster model
Parquet file formats while ensuring the loading and where data is stored across multiple nodes organized into data
transformation are carried out for their future processing. This centers. This model ensures high fault tolerance as data is
diagram provides an overview of how Spark is going to intake replicated across different nodes and geographically
data in various formats, perform processing based on distributed data centers.
distributed computation mechanisms, and move the outcome Redis is a key-value database known for its high operation
to other systems or storage devices for further analysis. It gives speed and is used for caching, real-time data processing, and
other tasks requiring minimal latency (figure 5).
191
 Eduard Shaikhulov, International Journal of Advanced Trends in Computer Science and Engineering, 14(4), July – August 2025, 190 - 193

One of the most important optimization mechanisms is

indexing, which accelerates query execution by structuring
data in advance. In relational DBMS, indexing is based on
B-trees, hash tables, or bitmap indexes that reduce the number
of read operations during record searches [6]. In NoSQL
databases like MongoDB, the creation of indexes on document
fields significantly accelerates data retrieval. However,
excessive indexing increases the cost of updating data, and
Figure 5: Redis architecture diagram therefore requires a balance between search speed and
indexing overhead.
Redis will play nicely in all server application architectures, Another optimization technique is data partitioning, which
providing a frontend-backend-worker nodes caching layer to includes the division of tables or collections into logical
reduce the latency of the primary database, thus speeding up segments called partitions. This, in turn, enables parallel data
the data processing and system performance in general. processing, hence increasing query execution speed. As such,
Another distinctive feature of Redis is the in-memory there exist a variety of partitioning techniques: horizontal,
approach to data processing; it enables the possibility of an vertical, and by range. In the context of NoSQL databases,
immediate response in really highly loaded systems. such as Cassandra, for instance, partitioning forms the basis
Relational DBMS, such as SQL Server, Oracle, and for scalability, since this will ensure even distribution of load
PostgreSQL, are also adapted to handle big data through the across nodes.
implementation of specialized technologies (table 1). Handling big data also means efficient management of data
loading. In real-time systems, stream processing is applied
Table 1: Relational DBMS and their work with big data [5] where new records are loaded upon arrival without
DBMS Big data processing Usage features overloading the primary database. Traditional relational
methods DBMS often use batch processing, where data is written to the
SQL Server Integration with Hadoop Used in system in large blocks, reducing overhead costs.
and Spark; support for corporate BI Asynchronous processing and intermediate result caching
analytics and machine systems and further contribute to reducing system response time.
learning. analytics. Data compression, in this regard, reduces the volume of stored
Oracle Exadata – an optimized Designed for information considerably, reduces the actual need for storage,
storage system for high-load and improves the overall system throughput accordingly [7].
high-performance financial and Various compression algorithms, such as LZ4, Snappy, and
computing. cloud solutions. Zstandard, are applied in the storage and transportation of data
with minimal loss of performance. For example, columnar
PostgreSQL JSON support, parallel Popular among
compression is followed in systems dealing with textual and
query execution, open-source
semi-structured data to reduce redundancy and speed up
integration with external developers and
analytics.
data sources. analytical
Moreover, deduplication in highly informed databases helps in
systems.
the elimination of duplicate records to save storage; this is
applicable on both application and database levels, which
Each of these technologies has its specific tasks to perform;
again would find relevance with high data redundancy
collectively, they make the ecosystem of big data. Their
systems, logging, monitoring, and analytics platforms.
correct application provides the necessary efficiencies in
processing data, extracting insights, and building analytic
5. SECURITY AND DATA MANAGEMENT
models applied in sectors such as finance, healthcare,
manufacturing, and science.
With the unparalleled increase in volumes of data and their
huge usage in nearly every other industry, the security and
4. OPTIMIZATION OF DATABASES FOR BIG DATA
management of information become a highly crucial task for
PROCESSING
the developers and administrators of the database management
system (table 2).
As the amount of information handled by today's databases is
growing significantly, efficient optimization methods must be
found to increase their performance and minimize query
processing delay times. The optimization techniques may
range from indexing, partitioning, and data load management
to compression and deduplication methods.

192
 Eduard Shaikhulov, International Journal of Advanced Trends in Computer Science and Engineering, 14(4), July – August 2025, 190 - 193

Table 2: Methods for ensuring security and managing big data [8] The modern concept of database protection involves
Method Description Application encryption, access control, monitoring, and auditing in order
Authentication Mechanisms for Used in corporate to minimize risks related to data disclosure and unauthorized
and verifying user and cloud access. Big data management is to be provided in conformity
authorization identity and databases. with legal requirements and to ensure the fault tolerance of
restricting access systems. Future development will include enhancing data
rights. processing methodologies, integration of artificial intelligence
Data Encoding Applied for data tools, and further development of analytical and visualization
encryption information using protection in capabilities. These solutions let organizations unlock the
cryptographic storage and complete power of big data, thus making big data a prime asset
algorithms. transmission. in every sector.
Access control Restricting user Implemented
rights to view, through REFERENCES
modify, and delete role-based access 1. A. O. Adewusi, O. F. Asuzu. Business intelligence in the
data. control (RBAC, era of big data: a review of analytical tools and
ABAC). competitive advantage, Computer Science & IT
Monitoring Tracking user actions Used for Research Journal, Vol 5, no. 2, pp. 415-431, 2024.
and auditing and detecting identifying 2. A. Malikov. Digital transformation and its impact on
anomalies in system suspicious the structure and efficiency of modern business,
operations. activity. Annali d’Italia, no. 62, pp. 112-115, 2024.
Backup and Creating copies of Applied in 3. L. G. Ahmad. Spark, Hadoop, and Beyond: Exploring
recovery data for restoration in high-load and Distributed Frameworks for Metaheuristic
case of failure. critical systems. Optimization in Big Data Analytics. 2024.
Data masking Concealing parts of Used in testing 4. V. H. Olivera, G. Ruizhe, R. C. Huacarpuma, P. B. Silva,
information to and analytical A. M. Mariano, M. Holanda. Data modeling and
protect sensitive environments. NoSQL databases-a systematic mapping review, ACM
data. Computing Surveys (CSUR), Vol. 54, no. 6, pp. 1-26,
2021.
Big data security and its efficient management cannot be 5. B. El Idrissi, S. Baïna, A. Mamouny, M. Elmaallam.
ensured without considering information protection at all RDF/OWL storage and management in relational
levels, starting from user authentication to activity monitoring database management systems: A comparative study,
and data leak prevention. In this case, the consideration is not Journal of King Saud University-Computer and
limited to the technological aspects of security; the regulatory Information Sciences, Vol. 34, no. 9, pp. 7604-7620,
requirements around personal data processing, such as GDPR 2022.
(General Data Protection Regulation), CCPA (California 6. M. M. Rahman, I. Siful, Md Kamruzzaman, H. J. Zihad.
Consumer Privacy Act), and HIPAA (Health Insurance Advanced query optimization in SQL databases for
Portability and Accountability Act), should also be taken into real-time big data analytics, Academic Journal on
consideration. It is relevant that a data management system be Business Administration, Innovation & Sustainability,
adapted to the distributed storage architecture, ensuring fault Vol. 4, no. 3, pp. 1-14, 2024.
tolerance with minimal losses in the case of any failure. 7. L. Dinesh, K. G. Devi. An efficient hybrid optimization
of ETL process in data warehouse of cloud
6. CONCLUSION architecture, Journal of Cloud Computing, Vol. 13, no.
1, pp. 122024.
Big data and the related database technologies are instruments 8. A. Israfilov. Stages of development and
of the fundamental class of a modern digital world that ranges implementation of information security policies in
from real-time analyses to machine learning. Working with big organizations, Danish scientific journal, no. 90, pp.
data requires special architectures, such as implementations of 131-134, 2024.
distributed computing, NoSQL solutions, or adapted relational
DBMS. Among the effective ways for database optimization, a
distinction is made for indexing, partitioning, compressing,
and managing data load. These methods ensure the high
performance, scalability, and reliability of storage and
processing systems and are a core in developing digital
platforms. Security and data management also remain crucial
issues.

193