Performance Enhancement of Blockchain System Using Dynamic

Sharding Technique for Marks Card Management

Swapnil M Maladkar, Praveen M Dhulavvagol and S G Totad

School of Computer Science and Engineering, KLE Technological University, Hubli, 580031, India

Keywords: Blockchain, Scalability, Dynamic, Sharding, Partition

Abstract: Blockchain technology, while revolutionary in its approach to secure and decentralized data management,

faces significant scalability challenges. As transaction volumes increase, traditional blockchain networks

often struggle with performance bottlenecks, leading to slower transaction times, higher costs, and increased

computational demands. This paper presents a novel solution to these scalability issues through the

implementation of a dynamic sharding algorithm. The proposed algorithm introduces a flexible framework

for partitioning a blockchain network into multiple dynamically managed shards. Each shard operates as an

independent blockchain, allowing the system to distribute transaction loads more effectively and adapt to

varying levels of activity. To demonstrate the effectiveness of this approach, we apply the dynamic sharding

algorithm to a blockchain-based student marks card management system. This application benefits from

improved scalability, security, and efficiency, addressing the limitations of traditional centralized record-

keeping methods. The results show that dynamic sharding achieves a 19.5% increase in transaction throughput

and a 25% reduction in latency, while also maintaining the integrity and transparency of the blockchain

network.

1 INTRODUCTION

A blockchain is a decentralized ledger consisting of

an expanding series of records called blocks, which

are securely interconnected using cryptographic

hashes. Each block contains details of the preceding

block, creating a sequential chain where each new

block links to its predecessor. As a result, blockchain

transactions are immutable; once recorded, the data

within a block cannot be modified without changing

all subsequent blocks. Blockchains are usually

maintained by a peer-to-peer (P2P) network,

functioning as a public distributed ledger, where

nodes collectively follow a consensus algorithm to

validate and add new transaction blocks (Blockchain,

2024). Blockchain technology functions as a

decentralized database maintained by all participating

nodes, which boosts the reliability of Trusted Third-

Party Auditors (TPAs) and enhances the security of

data auditing processes. Its capacity to generate an

immutable record of transactions ensures the integrity

of data and auditing outcomes, making it challenging

for malicious actors to alter the information stored on

the blockchain (Y. Miao, 2024).

Despite its many advantages, blockchain technology

faces significant scalability challenges. Unlike

centralized systems that can quickly access their user

databases, blockchain systems struggle with

efficiently handling large volumes of information,

resulting in lower scalability (Hisseine, 2022).

Scalability in blockchain technology depends on

various factors, including transactions per second,

block size, chain size, and digital signatures. As the

number of transactions increases, several issues arise:

the average confirmation time for a transaction

increases, network transaction fees rise, the difficulty

of mining blocks escalates, and consequently, the

required computational power and resources also

grow. Additionally, the block size increases. As a

result, the system becomes slower, more expensive,

and unsustainable. Due to these challenges,

scalability has become a key focus area for

researchers in the blockchain field (Hemlata Kohad,

2020).

The conventional approach of managing student

marks cards in educational institutions typically rely

on centralized databases and physical records. These

systems involve manual entry of student marks,

storage in centralized servers, and physical

582

Maladkar, S. M., Dhulavvagol, P. M. and Totad, S. G.

Performance Enhancement of Blockchain System Using Dynamic Sharding Technique for Marks Card Management.

DOI: 10.5220/0013582100004664

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 3rd International Conference on Futuristic Technology (INCOFT 2025) - Volume 1, pages 582-591

ISBN: 978-989-758-763-4

documentation, which can lead to several limitations.

Firstly, centralized systems are vulnerable to single

points of failure, making them susceptible to data

breaches, hacking, and system downtimes. Secondly,

manual data entry is prone to human errors, which can

lead to inaccuracies in student records. Additionally,

centralized databases can be tampered with, leading

to potential data manipulation or unauthorized access.

Physical records, on the other hand, are at risk of

damage, loss, and degradation over time.

To address this challenge, we propose a novel and

comprehensive solution that utilizes dynamic shard

generation technique. The main goal is to enhance the

scalability of blockchain network while maintaining

the fundamental principles of decentralization,

security, transparency, and immutability (Tennakoon,

2022). Dynamic sharding algorithm, dynamically

creates and manages multiple shards within the

blockchain network, with each shard operating as an

independent blockchain, complete with its own set of

verified users and data. By efficiently distributing the

transaction load across these dynamically adjusted

shards, the approach optimizes resource utilization

and enhances overall system performance, thereby

enabling seamless scalability (Wang, 2019). In the

context of Record Management applications, this

method can manage and scale various aspects such as

student marks card management, ensuring efficient,

secure, and scalable data handling.

Blockchain technology offers a robust solution to

these limitations by providing a decentralized,

tamper-proof, and transparent system for managing

student marks cards. By utilizing a distributed ledger,

each student's marks can be recorded as a transaction

in an immutable ledger, ensuring data integrity and

authenticity. The decentralized nature of blockchain

eliminates the single point of failure, enhancing the

security and reliability of the system. Moreover, the

transparency of blockchain allows for easy

verification and auditing of records, ensuring that any

attempt to alter or manipulate data is easily detectable

(I.Mohammed Ali, 2021). The use of cryptographic

techniques, such as public key and private key

encryption, further secures student data. Each student

and institution can have their own pair of public and

private keys. The public key is used to encrypt the

data, making it accessible only to those with the

corresponding private key, ensuring that only

authorized parties can access or modify the data. This

provides enhanced security measures against

unauthorized access. By implementing blockchain in

student marks card management, educational

institutions can achieve a more secure, accurate, and

efficient system for storing and managing student

records, leveraging the power of distributed ledgers

and cryptographic security to protect student data.

Dynamic shard generation represents an effective

and integrated strategy to address scalability issues in

blockchain networks. Our approach focuses on

increasing transaction throughput, minimizing

network latency, and optimizing resource utilization

to facilitate the wider adoption of blockchain

technology in various industries. This study's main

objective is to implement sharding within a

blockchain network to boost transaction throughput

and shorten transaction confirmation times. By

distributing the processing load across multiple

shards, we aim to enhance the network's capacity to

handle transactions more efficiently and rapidly.

A key contribution to enhancing the scalability of

blockchain networks is the introduction of a dynamic

sharding algorithm. This algorithm partitions the

network into multiple shards, each operating as an

independent blockchain with its own verified users

and data. By effectively distributing transaction loads

across these dynamically managed shards, the

algorithm enhances resource utilization and boosts

overall system performance.

The paper is organized with discussions on

Related work in Section 2, followed by the Proposed

Methodology of the shard generation algorithm in

Section 3, and Results and Analysis in Section 4, A

comparison with existing solutions highlights the

advantages and effectiveness of the proposed

approach.

2 RELATED WORK

This section offers an extensive review of the current

literature and research on blockchain scalability,

highlighting the field's status and the different

strategies proposed to address scalability challenges.

However, traditional distributed databases and

blockchain systems each display unique failure

modes, as noted in (Praveen M Dhulavvagol, 2023).

Darllaine R. et al. (Darllaine R, 2024) discussed

the increasing popularity of blockchain technology

but noted its significant scalability challenges,

particularly in public blockchain platforms like

Bitcoin and Ethereum. The main issues include low

throughput, high transaction latency, and high energy

consumption. This paper explores various state-of-

the-art solutions, categorizing them into three layers.

Layer 0 focuses on improving network information

dissemination through propagation protocols. Layer 2

includes on-chain solutions like redesigning block

structures, implementing Directed Acyclic Graphs

Performance Enhancement of Blockchain System Using Dynamic Sharding Technique for Marks Card Management

583

(DAG), sharding techniques, and Segregated Witness

(SegWit). Layer 3 comprises off-chain solutions like

side-chain techniques, payment channels, and cross-

chain techniques. Sharding emerges as a notable

approach, with solutions like Ostraka achieving high

TPS but struggling with bandwidth and

computational resource sharding, and RapidChain

performing well but being prone to partitioning

attacks. The paper concludes that while many

promising solutions exist, each has limitations, and

future research should focus on integrating offline

channels with the Lightning Network approach for

applications beyond payment channels.

Khacef et al. (Khacef, 2021) proposed SecuSca, a

novel approach to address the trade-off between

security and scalability in blockchain systems.

Blockchain technology, particularly in public

blockchains like Bitcoin and Ethereum, faces

scalability issues due to the full replication of the

entire blockchain on all nodes, leading to storage and

performance problems. SecuSca introduces a

dynamic sharding mechanism that reduces storage

load by decreasing block replication across the

network. It retains block headers but removes

transaction data from older blocks on most nodes,

allowing for more efficient storage use. New blocks

are initially replicated across many nodes for security,

but as they get buried deeper in the chain, replication

is reduced while maintaining headers. An

optimization function, R(d), determines replication

levels based on block depth, balancing security and

scalability with parameters α and γ. Simulations

comparing traditional full replication and SecuSca

showed significant reductions in storage

requirements while maintaining security. SecuSca

allows the blockchain to store more transactions with

the same total storage capacity, dynamically adjusting

block replication based on age and depth. Future work

will focus on improving transaction verification and

developing inter-shard communication protocols.

SecuSca offers a promising solution to blockchain

scalability by optimizing storage use, maintaining

security, and increasing transaction capacity.

Qinglin Yang et al. (Qinglin Yang, 2024) offers

an in-depth overview of blockchain sharding, a

technique aimed at enhancing blockchain scalability

without compromising decentralization. Sharding

divides consensus nodes into smaller groups,

allowing parallel processing of transactions and smart

contracts, which significantly improves throughput

and reduces transaction confirmation latency. The

document categorizes sharding techniques into

several key areas: processing cross-shard transactions

with methods like ByShard and BrokerChain;

balancing shard workloads using approaches such as

TxAllo and LB-Chain; efficient smart contract

execution through solutions like Jenga and Prophet;

shard reorganization with techniques like S-Store;

enhancing security via multi-shard oversight with

CoChain; accelerating block confirmation; and

stabilizing transaction pools. Experimental results

highlight the superiority of sharding protocols like

Monoxide and Metis over single-shard systems.

Despite its benefits, sharding faces challenges such as

cross-shard communication, shard rebalancing,

security concerns, implementation complexity, and

smart contract compatibility. The article encourages

further research to address these challenges and fully

harness sharding's potential to improve blockchain

scalability.

Amiri et al. (Amiri, 2019) introduces a model for

sharding permissioned blockchains to significantly

enhance scalability by leveraging Byzantine fault-

tolerant protocols among identified nodes, which

traditionally require 3f+1 nodes to tolerate f failures.

The proposed model innovatively optimizes the use

of surplus nodes by partitioning them into clusters

and sharding data across these clusters, each

containing 3f+1 nodes. The system supports two

types of transactions: intra-shard, which occur within

a single shard, and cross-shard, which span multiple

shards. It generalizes the blockchain ledger from a

linear chain to a Directed Acyclic Graph (DAG),

where each block contains a single transaction to

boost performance. Intra-shard transactions are

ordered within their cluster, while cross-shard

transactions create links between shards in the DAG,

with each cluster maintaining its view of the ledger.

Key features of this model include parallel

processing, allowing different clusters to handle

independent transactions simultaneously; scalability,

achieved by adding more clusters and shards; and

efficient resource utilization, employing extra nodes

in separate clusters. Challenges highlighted include

designing consensus protocols for both intra-shard

and cross-shard transactions, balancing the load

across shards to maximize parallelism, and ensuring

security and consistency throughout the sharded

system. The potential benefits of this model are

substantial, offering improved throughput through

parallel processing, better scalability compared to

non-sharded blockchains, and more efficient resource

use in large permissioned blockchain networks. This

theoretical model and architecture for sharded

permissioned blockchains set the stage for future

implementation and protocol design, aiming to

address the scalability limitations of current

INCOFT 2025 - International Conference on Futuristic Technology

584

blockchain systems and leverage additional nodes for

parallel processing.

3 PROPOSED METHODOLOGY

The proposed dynamic sharding approach

significantly enhance blockchain scalability and

performance, with a specific focus on its application

in student marks card management systems. The use

of cryptographic techniques, such as public key and

private key encryption (Ahmad, 2023), further

secures student data. Each student and institution can

have their own pair of public and private keys. The

public key is used to encrypt the data, making it

accessible only to those with the corresponding

private key, ensuring that only authorized parties can

access or modify the data. This provides enhanced

security measures against unauthorized access. By

implementing blockchain in student marks card

management, educational institutions can achieve a

more secure, accurate, and efficient system for storing

and managing student records, leveraging the power

of distributed ledgers and cryptographic security to

protect student data.

Traditional static partitioning methods often fail

to adapt to fluctuating network conditions and

transaction volumes, leading to inefficiencies and

suboptimal performance. The proposed methodology

introduces a dynamic adjustment mechanism that

allows for real-time reconfiguration of partition sizes

and locations, responding adaptively to changing

demands. This approach not only improves resource

utilization by balancing the load across partitions but

also enhances transaction throughput and minimizes

latency. By leveraging dynamic sharding, we aim to

address common scalability challenges and ensure

robust data integrity in a decentralized environment.

The proposed system will enable a more responsive

and efficient blockchain infrastructure, capable of

handling varying workloads associated with student

records management. This innovation promises to

overcome the limitations of static sharding

approaches, providing a scalable and adaptable

solution that meets the evolving needs of educational

institutions and their data management requirements.

The shard generation algorithm mainly consists of

8 steps as follows:

3.1 Initialization:

The dynamic shard generation algorithm begins by

setting up the foundational parameters necessary for

efficient shard management. Initially, the system

defines the maximum number of partitions, or shards,

that the blockchain network can support. This

maximum is determined based on the expected

scalability needs and the resources available.

Partitions are then initialized, either based on

historical data, equal distribution, or an initial

workload assessment. These partitions are the

segments of the blockchain that will handle

transactions and data independently. Additionally,

thresholds for load and transaction volume are

established.

3.2 Monitoring:

Once the partitions are in place, the system

continuously monitors their performance in real-time.

This involves tracking various metrics such as the

load on each partition (including CPU usage, memory

consumption, and I/O operations) and the transaction

volume being processed. Real-time data collection

helps in understanding the current state of each

partition and provides a basis for making necessary

adjustments. By calculating average load and

transaction volume across all partitions, the system

establishes a baseline against which individual

partitions can be assessed.

3.3 Analysis:

With real-time monitoring data in hand, the algorithm

performs an analysis to identify performance issues.

Partitions are evaluated against predefined

thresholds. Those that exceed the load and transaction

volume thresholds are classified as overloaded, while

those significantly below half of these thresholds are

identified as underutilized. This analysis highlights

partitions that require intervention, either to

redistribute workload or to consolidate resources.

3.4 Adjustment:

To address overloaded partitions, the system

identifies adjacent or less loaded partitions that can

absorb some of the excess load. Data and transactions

are redistributed from the overloaded partitions to

these selected partitions, which helps in balancing the

overall load. Conversely, underutilized partitions are

evaluated for potential merging opportunities.

Merging involves combining underutilized partitions

into a larger partition to improve efficiency and

reduce overhead.

Performance Enhancement of Blockchain System Using Dynamic Sharding Technique for Marks Card Management

585

3.5 Reconfiguration:

After adjustments, the system must update its

metadata and routing tables to reflect the new

partition configurations. This ensures that all network

nodes are aware of the changes and can correctly

route transactions and access data have based on the

updated partition structure. This reconfiguration step

is essential for maintaining the coherence and

functionality of the blockchain network.

3.6 Validation:

Following reconfiguration, the system performs

validation checks to ensure that data integrity and

consistency are maintained. This involves verifying

that no data has been lost or corrupted during the

partition adjustments. Additionally, performance

metrics such as transaction throughput, latency, and

resource utilization are assessed to confirm that the

adjustments have achieved the desired improvements.

3.7 Iteration:

The dynamic sharding process is iterative, meaning

that monitoring, analysis, and adjustment are

continuous activities. The system regularly repeats

these steps to adapt to changing conditions and

workload patterns. By incorporating adaptive

strategies and predictive analytics, the system can

anticipate future load patterns and proactively adjust

partitions, maintaining optimal performance and

scalability over time.

3.8 Scalability and Expansion:

As the blockchain network grows, the system may

need to expand the shard pool to accommodate

increasing data and transaction volumes. This

involves dynamically adding new shards and

adjusting partitioning strategies to integrate these new

shards effectively. The scalability and expansion

process ensure that the network remains responsive

and efficient as it scales up to handle more extensive

datasets and transaction loads.

Figure 1. Represents architecture of dynamic

sharding architecture that enhances blockchain

scalability and performance, specifically for student

record management systems. It is divided into four

phases: Initialization, Monitoring, Analysis, and

Adjustment. In the Initialization phase, the maximum

number of partitions is set, and partitions are

initialized with predefined load and transaction

thresholds. The Monitoring phase involves real-time

data collection on performance metrics like

transaction throughput (TPS), latency, and resource

Figure 1: Proposed dynamic sharding system architecture

utilization. During the Analysis phase, this data is

evaluated against thresholds to classify partitions as

overloaded or underutilized. Finally, in the

Adjustment phase, load is dynamically redistributed

from overloaded partitions to balance the network,

and underutilized partitions are merged to optimize

resource use. This architecture ensures efficient,

scalable blockchain operations by adapting to real-

time network demands, leading to balanced resource

utilization, improved TPS, and reduced latency.

Parameters involved Dynamic Shard Generation:

Maximum Number of Partitions (P

max

): The

maximum number of partitions, denoted as P

max

defines the upper limit of shards that the blockchain

network can support. This parameter is crucial as it

determines the scalability potential of the system,

setting a cap on how many parallel partitions can

exist. By establishing P

max

, the system ensures

controlled growth and prevents resource

overcommitment. It balances the need for increased

capacity with the practical limitations of

computational and storage resources, thus enabling

efficient and sustainable expansion.

Initial Partitions (P): Initial partitions,

represented as P= {P

, P

, …, P

}, are the starting set

of shards in the blockchain network. These partitions

are established based on historical data, equal

distribution strategies, or initial workload

assessments. Proper initialization is essential for

ensuring a balanced distribution of data and

transactions from the outset. It provides a

foundational structure upon which dynamic

adjustments can be made, ensuring that the system

operates effectively and efficiently right from the

beginning.

Load Threshold (T

load

): The load threshold, T

load

, specifies the maximum acceptable load for each

partition, encompassing factors such as CPU usage,

INCOFT 2025 - International Conference on Futuristic Technology

586

memory consumption, and I/O operations. When a

partition’s load exceeds T

load

, it signals that the

partition is under significant strain and may require

intervention. This parameter helps in proactively

managing performance by identifying partitions that

are overloaded, enabling timely adjustments to

redistribute the workload and prevent potential

system bottlenecks.

Transaction Volume Threshold (T

): Transaction

volume threshold, T

, sets the upper limit for the

number of transactions that each partition can handle

effectively. This threshold is critical for maintaining

system performance and responsiveness by

preventing partitions from being overwhelmed by

high transaction volumes. When a partition’s

transaction volume surpasses T

, it triggers actions to

balance the load, ensuring that transaction processing

remains efficient and delays are minimized.

Average Load (avg

load

): The average load, avg

load

, represents the mean load across all partitions in the

network. Calculating this average provides a

benchmark for evaluating the performance and load

distribution within the system. It helps in identifying

partitions that deviate significantly from the norm,

allowing for targeted adjustments to balance the load.

This parameter is essential for maintaining overall

system performance and ensuring that no single

partition becomes a bottleneck.

Average Transaction Volume (avg

): The

average transaction volume, avg

, indicates the mean

number of transactions processed by each partition.

By calculating this average, the system establishes a

baseline for typical transaction activity, aiding in the

detection of partitions with abnormal transaction

levels. Monitoring avg

is crucial for ensuring that

transaction loads are evenly distributed and that no

partition is overwhelmed, thereby optimizing

processing efficiency.

Data Redistribution Parameters: Data

redistribution parameters define the criteria and

methods for transferring data between partitions.

These include the amount of data to be moved, the

frequency of redistribution, and the techniques used

for transferring data. Properly defining these

parameters is crucial for maintaining balance among

partitions and ensuring that workload redistribution is

carried out smoothly. Effective data redistribution

helps in optimizing resource utilization and

preventing performance degradation due to uneven

data distribution.

Merge Criteria: Merge criteria are the conditions

under which underutilized partitions are combined to

improve efficiency. This includes the selection

process for which partitions to merge and the rules for

consolidating data and resources. By establishing

clear merge criteria, the system ensures that partition

consolidation is done in a way that maximizes

resource utilization and minimizes overhead. This

parameter is key for reducing system complexity and

improving overall performance by consolidating

underused resources.

Consistency and Integrity Checks: Consistency

and integrity checks involve mechanisms to ensure

that data remains accurate and reliable during and

after partition adjustments. This includes verifying

that no data is lost or corrupted during data

redistribution and merging processes. Implementing

these checks is vital for maintaining the

trustworthiness of the blockchain network, ensuring

that all data remains intact and that system operations

proceed without disruptions.

Reconfiguration Metadata: Reconfiguration

metadata encompasses the data and routing

information that must be updated to reflect new

partition configurations. This includes changes to

partition addresses, locations, and boundaries.

Accurate updating of reconfiguration metadata is

essential for ensuring that all network nodes are

aware of the new partition structure and can properly

route transactions and access data. This parameter is

crucial for maintaining seamless operation and

coherence across the blockchain network after

adjustments are made.

The procedure for the dynamic hybrid sharding

algorithm involves several key steps to ensure

efficient load balancing and transaction management

across blockchain partitions. Initially, the system

continuously monitors the current load and

transaction volume for each partition, gathering real-

time data to inform adjustments. The average load

(avg

load

) and average transaction volume (avg

) are

computed across all partitions to establish a baseline

for comparison. Partitions are then categorized into

overloaded and underutilized based on their

performance relative to the averages. For overloaded

partitions, the algorithm transfers a portion of the data

to adjacent or least-loaded partitions, thereby

alleviating the excessive burden. Conversely, for

underutilized partitions, the system evaluates the

potential benefits of merging them with adjacent or

related partitions and adjusts boundaries accordingly.

This redistribution ensures that load and transaction

values are balanced. These steps—monitoring,

analysis, and adjustment—are repeated at regular

intervals or whenever significant changes in load or

transaction volume are detected, ensuring the system

remains responsive to fluctuations. The updated

partition assignments, reflecting these dynamic

Performance Enhancement of Blockchain System Using Dynamic Sharding Technique for Marks Card Management

587

adjustments, are then returned as the final output,

maintaining an optimal and efficient blockchain

environment.

Dynamic shard generation algorithm:

Algorithm 1: Adjust Partitions

Input:

1. Partitions

2. Load_data

3. tx_data

4. T_load

5. T_tx

Output: Partiton adjustments made to balance load

and transaction volume.

Procedure:

# Calculate the average load and transaction volume across

all partitions.

avg_load = sum(load_data.values()) / len(partitions)

avg_tx = sum(tx_data.values()) / len(partitions)

#Identify overloaded partitions and underutilized partitions.

overloaded = [p for p in partitions if load_data[p] > T_load

or tx_data[p] > T_tx]

underutilized = [p for p in partitions if load_data[p] <

T_load / 2 or tx_data[p] < T_tx / 2]

#Call redistribute_load(overloaded_partition,

target_partitions, load_data, tx_data) to redistribute data

and #balance load

for p in overloaded:

target_partitions =

find_adjacent_or_least_loaded(partitions, load_data,

tx_data)

redistribute_load(p, target_partitions, load_data, tx_data)

#Call merge_or_adjust(underutilized_partition, partitions,

load_data, tx_data) to either merge or adjust the #partition

boundaries.

for p in underutilized:

merge_or_adjust(p, partitions, load_data, tx_data)

return partitions

Algorithm 2: Redistribute load across partitions.

Input:

1. overloaded_partition

2. target_partitions

3. load_data

4. tx_data

Output: The load and transaction volume of

overloaded partition is redistributed across the target

partitions

Procedure:

# Transfer data to target partitions

#Distribute the load and transaction volume from the

overloaded partition to target partitions.

for target in target_partitions:

transfer_data(overloaded_partition, target)

load_data[target] += load_data[overloaded_partition] /

len(target_partitions)

tx_data[target] += tx_data[overloaded_partition] /

len(target_partitions)

load_data[overloaded_partition] =

get_current_load(overloaded_partition)

tx_data[overloaded_partition] =

get_current_tx_volume(overloaded_partition)

Algorithm 3: merge_or_adjust(underutilized_partition,

partitions, load_data, tx_data):

Input:

1. underutilized_partition

2. partitions

3. load_data

4. tx_data

Output:

1. modify the list of partition if a merge operation

occurs.

2. updates the load_data and tx_data

Procedure:

#Identify adjacent partitions to the underutilized partition.

adjacent_partitions =

find_adjacent_partitions(underutilized_partition,

partitions)

#If merging is feasible, perform the merge operation.

if can_merge(underutilized_partition, adjacent_partitions):

merge_partitions(underutilized_partition,

adjacent_partitions)

#If merging is not feasible, adjust the boundaries of the

underutilized partition.

else:

adjust_boundaries(underutilized_partition,

adjacent_partitions, load_data, tx_data)

4 RESULTS AND ANALYSIS

In evaluating the effectiveness of our proposed

dynamic sharding algorithm for blockchain

scalability within a student record management

context, we conducted a series of experiments using

a simulated blockchain network. The results

demonstrated significant improvements across

several key metrics.

Firstly, in terms of transaction throughput, the

initial static partitioning setup achieved an average of

241 transactions per second (TPS). After

implementing our dynamic sharding algorithm, the

TPS increased to 288, representing a substantial

improvement of approximately 19.5% as shown in

Fig 2. This enhancement highlights the algorithm's

capability to handle higher transaction volumes

through dynamic adjustments. Regarding latency, the

initial static partitioning setup exhibited an average

transaction confirmation time of 2.688 seconds. With

dynamic sharding, this latency decreased to an

average of 2.014 seconds, marking a reduction of

approximately 25% which is represented in Fig 3.

INCOFT 2025 - International Conference on Futuristic Technology

588

This improvement indicates that dynamic partitioning

allows for quicker transaction confirmations by

preventing any single partition from becoming a

bottleneck.

Figure 2: Comparison of Transaction Throughput (TPS)

Figure 3: Comparison of Latency

Table 1. Comparison of Transaction Throughput and

Latency.

No. of

Transactio

Transactions

Per Second

(

TPS

)

Latency

(seconds)

Initia

static

setup

Dynami

Initia

static

Setu

Dynami

1 100 68 84 1.73 1.21

2 150 113 131 1.94 1.36

3 200 132 179 2.13 1.57

4 250 178 202 2.39 1.63

5 300 211 259 2.42 1.87

6 350 279 304 2.61 2.06

7 400 283 351 2.84 2.29

8 450 331 423 3.29 2.41

9 500 394 447 3.57 2.69

10 550 422 509 3.96 3.05

In terms of resource utilization, the initial static

partitioning setup saw uneven distribution, with some

nodes being overutilized while others remained

underutilized. After applying the dynamic sharding

algorithm, resource utilization became more

balanced, with CPU and memory usage evenly

distributed across all nodes. This balance

demonstrates the algorithm's effectiveness in

distributing the workload and preventing resource

exhaustion on any single node. Load distribution also

saw significant improvements. Initially, load

distribution was skewed, with some partitions

frequently overloaded while others were

underutilized. The dynamic sharding approach

resulted in a much more uniform load distribution,

with partitions dynamically resized and reallocated

based on real-time network conditions. This

uniformity helped maintain optimal performance and

prevented overloading. Lastly, data integrity was

consistently maintained throughout all experiments,

with no incidents of data loss or corruption observed

in both setups. This consistency confirms that the

dynamic adjustments do not compromise the

correctness and consistency of the blockchain data,

ensuring reliable data management for student marks.

Figure 4: CPU Usage Comparison Across Nodes

Figure 5: Memory Usage Comparison Across Nodes

Fig 4. illustrates the CPU usage across different

nodes in the blockchain network, comparing the

Performance Enhancement of Blockchain System Using Dynamic Sharding Technique for Marks Card Management

589

initial static partitioning setup with the dynamic

sharding setup. This approach achieves a more

balanced distribution of CPU usage across nodes,

addressing the inefficiencies observed in the initial

static partitioning setup.

Fig 5. shows the memory usage across different

nodes in the blockchain network, comparing the

initial static partitioning setup with the dynamic

sharding setup which results in a more balanced

distribution of memory usage across nodes,

mitigating the inefficiencies observed in the initial

static partitioning setup.

Figure 6: Load Distribution across Partitions

By comparing the initial static and dynamic load

distributions, the Fig 6. highlights the effectiveness of

the dynamic sharding algorithm in improving load

balancing across partitions.

In the context of student marks card management,

the dynamic sharding algorithm notably improved

system performance. The increased transaction

throughput and reduced latency ensured faster

updates and retrievals of student marks. The

enhanced resource utilization and balanced load

distribution contributed to smooth handling of high

request volumes, maintaining optimal performance

even under heavy load. By ensuring these

improvements while preserving data integrity, the

dynamic sharding approach significantly enhances

the scalability and responsiveness of the student

marks card management system.

5 CONCLUSIONS

The implementation of the dynamic sharding

algorithm has markedly enhanced blockchain

scalability in the context of student marks card

management. The algorithm achieved a notable

19.5% improvement in transaction throughput and a

25% reduction in latency, demonstrating its efficacy

in handling increased transaction volumes and

reducing confirmation times. Furthermore, the

dynamic approach led to more balanced resource

utilization and improved load distribution, ensuring

efficient performance across the network. These

advancements confirm the robustness and

effectiveness of dynamic sharding in optimizing

blockchain systems, making it a valuable

enhancement for decentralized applications

managing student records.

REFERENCES

Y. Miao, K. Gai, L. Zhu, K. -K. R. Choo and J. Vaidya,

"Blockchain-Based Shared Data Integrity Auditing and

Deduplication," in IEEE Transactions on Dependable

and Secure Computing, vol. 21, no. 4, pp. 3688-3703,

July-Aug. 2024, doi: 10.1109/TDSC.2023.3335413.

Blockchain - “https://en.wikipedia.org/wiki/Blockchain”,

2024.

Hisseine, M.A.; Chen, D.; Yang, X. The Application of

Blockchain in Social Media: A Systematic Literature

Review. Appl. Sci. 2022, 12, 6567.

https://doi.org/10.3390/app12136567.

Hemlata Kohad, Sunil Kumarand Asha Ambhaikar,

“Scalability Issues of Blockchain Technology”,

International Journal of Engineering and Advanced

Technology (IJEAT), vol. 9, no. 3, pp. 2385–2391, Feb.

2020, doi: 10.35940/ijeat.C5305.029320.

Tennakoon, Deepal & Gramoli, Vincent. (2022). Dynamic

Blockchain Sharding. 10.4230/OASIcs.FAB.2022.6.

Wang, Gang & Shi, Zhijie & Nixon, Mark & Han, Song.

(2019). SoK: Sharding on Blockchain. 41-61.

10.1145/3318041.3355457.

Praveen M Dhulavvagol, S G Totad, Performance

Enhancement of Distributed System Using HDFS

Federation and Sharding, Procedia Computer Science,

Volume 218, 2023, Pages 2830-2841, ISSN 1877-0509,

https://doi.org/10.1016/j.procs.2023.01.254.

Darllaine R. Lincopinis, Orven E. Llantos, the current

research status of solving blockchain scalability issue,

Procedia Computer Science, Volume 239, 2024, Pages

314-321, ISSN 1877-0509,

https://doi.org/10.1016/j.procs.2024.06.177.

Khacef, Kahina & Benbernou, Salima & Ouziri, Mourad &

Younas, Muhammad. (2021). Trade-Off Between

Security and Scalability in Blockchain Design: A

Dynamic Sharding Approach. 10.1007/978-3-030-

84337-3_7.

Xi, Jinwen & Zou, Shihong & Xu, Guosheng & Guo,

Yanhui & Lu, Yueming & xu, Jiuyun & Zhang,

Xuanwen. (2021). A Comprehensive Survey on

Sharding in Blockchains. Mobile Information Systems.

2021. 1-22. 10.1155/2021/5483243.

Qinglin Yang, Huawei Huang, Zhaokang Yin, et al. The

State-of-the-Art and Promising Future of Blockchain

Sharding. TechRxiv. January 26, 2024. DOI:

10.36227/techrxiv.170630359.94721690/v1

Amiri, Mohammad Javad & Agrawal, Divyakant &

Abbadi, Amr. (2019). On Sharding Permissioned

Blockchains. 10.1109/Blockchain.2019.00044.

INCOFT 2025 - International Conference on Futuristic Technology

590

Ahmad, Sartaj & Arya, Shubham & Gupta, Shobhit &

Singh, Puneeta & Dwivedi, Sanjeev. (2023). Study of

Cryptographic Techniques Adopted in Blockchain. 1-6.

10.1109/ICIEM59379.2023.10166591.

I.Mohammed Ali, Sura & Sharaf, Hussien. (2021). Using

Blockchain in University Management Systems.

Turkish Journal of Computer and Mathematics

Education (TURCOMAT). 12. 3305-3312.

10.17762/turcomat.v12i2.2389.

Dhulavvagol, Praveen M., S. G. Totad, and P. Pratheek.

"Enhancing Transaction Scalability of Blockchain

Network Using Sharding Technique." In International

Conference on Soft Computing for Security

Applications, pp. 253-269. Singapore: Springer Nature

Singapore, 2023.

Dhulavvagol, Praveen M., M. R. Prasad, Niranjan C.

Kundur, N. Jagadisha, and S. G. Totad. "Scalable

Blockchain Architecture: Leveraging Hybrid Shard

Generation and Data Partitioning." International

Journal of Advanced Computer Science and

Applications 14, no. 8 (2023).

Rane, Mukul & Singh, Shubham & Singh, Rohan &

Amarsinh, Vidhate. (2020). Integrity and Authenticity

of Academic Documents Using Blockchain Approach.

ITM Web of Conferences. 32. 03038.

10.1051/itmconf/20203203038.

M. Dhulavvagol Praveen, S G Totad, Mahadev Rashinkar,

Ribhav Ostwal, Suprita Patil, Priyanka M Hadapad,

Scalable Blockchain Architecture using off-chain IPFS

for Marks Card Validation, Procedia Computer

Science, Volume 215, 2022, Pages 370-379, ISSN

1877-0509,

https://doi.org/10.1016/j.procs.2022.12.039.

(https://www.sciencedirect.com/science/article/pii/S18

7705092202110X)

Surendran, K., L. Benny, and A. S. Mahesh. "Student

academic management system using blockchain

technology." Journal of Advanced Research in

Dynamical and Control Systems 12, no. 3 (2020):

1410-1415

Performance Enhancement of Blockchain System Using Dynamic Sharding Technique for Marks Card Management

591