SEBDA: A Secure and Efficient Blockchain Based Data Aggregation
Scheme
Sehrish Shafeeq and Mathias Fischer
Universität Hamburg, Germany
Keywords:
Secure Data Aggregation, Blockchain, End-to-End Integrity.
Abstract:
Data aggregation plays a vital role in collecting and summarizing data in the Internet of Things (IoT). Data ag-
gregation results are used to make a critical decision; therefore, the end-to-end integrity of the data aggregation
result is of utmost importance. Recently, blockchain-based data aggregation approaches have been proposed
that mainly focus on data confidentiality. However, existing approaches ignore two important requirements
1) the end-to-end integrity of data aggregation result and 2) the system’s scalability. This paper proposes a
blockchain-based data aggregation scheme to detect end-to-end integrity of aggregation results and identify
malicious aggregators. Furthermore, we improve the efficiency and scalability of the system by leveraging a
sidechain. Our simulation results indicate that the proposed system improves efficiency and scalability com-
pared to conventional blockchain-based data aggregation.
1 INTRODUCTION
The proliferation of the Internet of Things (IoT) is
driven by the promise to provide innovative solutions
to make life more efficient, integrated, and produc-
tive (Catarinucci et al., 2015; Jeong and Lee, 2014).
These solutions are equipped with numerous devices
that collect and exchange data to make intelligent de-
cisions. By 2023, the number of IoT devices is ex-
pected to reach 29.3 billion (Cisco, 2020). In many
scenarios, data consumers only require access to ag-
gregated data, rather than fine-grained data.
Data aggregation (Liu et al., 2019) is a process
of summarizing data obtained from multiple data
providers for statistical analysis, typically using a dis-
tributed approach called decentralized data aggrega-
tion. In this process, data providers send their raw
data to intermediate nodes that perform the aggrega-
tion function and then transmit the aggregation result
to the next intermediate node, continuing the process
until it reaches the querier. The intermediate nodes
are called aggregators.
Decentralized data aggregation reduces the num-
ber of data transmissions and improves scalability
(Krishnamachari et al., 2002). However, it has some
limitations. First, end-to-end integrity cannot be en-
forced easily, as malicious aggregators may not cor-
rectly perform aggregation operations and submit in-
correct aggregation results. End-to-end integrity im-
plies that the aggregation result is the output of the
correct computation of an aggregation function f per-
formed on the data values of a set of known data
providers. Second, the process lacks traceability and
transparency. The data submitted by data providers
and operations performed by aggregators are hidden
which makes difficult to find malicious aggregator.
In recent years, blockchain has attracted a lot of
attention for enhancing the security of IoT. However,
the current literature on the integration of blockchain
and data aggregation still faces some challenges. One
of the challenges is the scalability of the blockchain
network (Xie and Chen, 2021; Xie and Chen, 2021).
Furthermore, high transaction fees make microtrans-
actions infeasible and unattractive for data providers
to contribute to the process. Another issue is that ex-
isting approaches assume the aggregator to be "honest
but curious", meaning that it correctly performs com-
putation on the data which is not realistic assumption.
Therefore, there should be a mechanism that takes
advantage of blockchain technology in the data ag-
gregation process and simultaneously overcomes the
previously mentioned issues. This paper proposes a
decentralized data aggregation scheme with a 3-tier
architecture. The first tier comprises data providers
that send their data to sidechains representing the sec-
ond tier. A sidechain is a secondary blockchain that
helps to improve the system’s scalability. Sidechains
perform aggregation on the data values and submit ag-
Shafeeq, S. and Fischer, M.
SEBDA: A Secure and Efficient Blockchain Based Data Aggregation Scheme.
DOI: 10.5220/0012077900003555
In Proceedings of the 20th International Conference on Secur ity and Cryptography (SECRYPT 2023), pages 369-376
ISBN: 978-989-758-666-8; ISSN: 2184-7711
Copyright
c
2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)
369
gregation results to the mainchain. The mainchain
is the third tier and performs data aggregation on
the intermediate aggregation results submitted by the
sidechain. Following are the main contributions of the
proposed Secure and Efficient Blockchain based Data
Aggregation Scheme (SEBDA) scheme.
We propose a scheme that leverages the
blockchain and smart contracts for decentralized
data aggregation. The proposed scheme protects
the end-to-end integrity of the aggregation result
in the presence of malicious aggregators.
Our solution improves scalability by coupling on-
chain computation with sidechain computation.
The proposed scheme improves efficiency com-
pared to conventional blockchain-based data ag-
gregation.
The proposed scheme encourages honest partic-
ipation in the network by giving economic in-
centives and discourages malicious behaviors by
penalty mechanism.
The rest of the paper is organized as follows: Sec-
tion 2 summarizes the literature. Section 3 describes
the system model and the background. Section 4
presents the proposed solution. Section 5 provides se-
curity analysis and comparision with state-of-the-art
solutions. In Section 6, we performed performance
analysis. Finally, Section 7 concludes the paper.
2 RELATED WORK
Authors of (He et al., 2008) ensure integrity through
redundancy by constructing disjoint aggregation
trees. Each data provider sends its data to two ag-
gregation trees. The base station accepts aggrega-
tion results only if they agree with each other. In
(He et al., 2009), authors enhance (He et al., 2008)
with a method to identify malicious nodes. In this
scheme, data providers split their data into pieces and
send them to cluster members. The cluster head per-
forms data aggregation, and cluster members monitor
the aggregation result to identify malicious behaviour.
In (Li and Luo, 2012), authors proposed a homo-
morphic signature scheme to ensure the correctness of
the aggregation result. In (Leontiadis et al., 2015), au-
thors uses authentication tag to produce proof of cor-
rect aggregation computation. (Cui et al., 2018) pro-
posed a data aggregation scheme that achieves end-
to-end integrity using Homomorphic MAC (Agrawal
and Boneh, 2009). However, in this scheme, a single
private key is shared among all data providers for Ho-
momorphic MAC tag generation, which can be disas-
trous in the compromise of any node. (Vinodha and
Mary Anita, 2021) uses a concatenation based data
aggregation function which enables BS to recover in-
dividual data values and verifies integrity. However,
using concatenation-based data aggregation leads to
high communication costs.
(Fan et al., 2020) focus on removing a trusted
third party in smart grid data aggregation by using
blockchain. Each smart meter encrypts its data using
a Paillier cryptosystem (O’Keeffe, 2008) and sends it
to the mining node along with Boneh-Lynn-Shacham
short signature (Boneh et al., 2001). The mining
node aggregates the data, decrypts it and then pub-
lishes result on the blockchain. However, the mining
node is assumed to be honest but curious. (Loukil
et al., 2021) leverage smart contracts to group IoT de-
vices and choose an aggregator. Each data provider
first encrypts its data using the consumer public key
and then encrypts it using the aggregator public key.
However, double encryption increases computational
complexity for IoT devices. (Lu et al., 2021) pro-
pose a fault-tolerant secure data aggregation scheme.
Smart meters submit their data reports to Edge Server
(ES), which aggregates local data. ES then sends a
transaction that contains the aggregation result to a
master node for global data aggregation. (Xie and
Chen, 2021) leverages differential privacy and token-
based encryption to protect data confidentiality. In
this work, smart contracts perform data aggregation
and device identification. In summary, existing ag-
gregation schemes have limitations in terms of focus-
ing solely on data confidentiality, relying on a trusted
aggregator, and not considering scalability.
3 SYSTEM MODEL AND
BACKGROUND
3.1 System Model and Attacker Model
The system model consists of four entities: querier,
data providers, primary cluster head, and secondary
cluster head. The querier initiates the data aggrega-
tion process by issuing a query. The data provider
provides its data values
{
v
1
,... ,v
n
}
upon request of
the querier. Let D =
{
d
1
,. .. ,d
n
}
be a set of n data
providers, where each data provider has a unique
identifier. Each primary cluster head collects data val-
ues and performs intracluster data aggregation. The
data provider can also act as a primary cluster head.
Each secondary cluster head is also responsible for
performing aggregation function f
agg
and challenges
primary cluster head if it submits incorrect aggrega-
tion results.
The querier and data providers are assumed to be
honest. An adversary A can compromise a primary
SECRYPT 2023 - 20th International Conference on Security and Cryptography
370
cluster head or secondary cluster head. A compro-
mised primary cluster head or secondary cluster head
is called a malicious aggregator. A malicious aggre-
gator may selectively drop the data received from its
cluster members or manipulate the aggregate.
3.2 Background on Blockchain
Blockchain is a distributed ledger technology in
which a group of trustless actors works together to
maintain a record of transactions (Nakamoto, 2008).
The consensus is used to ensure that majority of the
network agree on a state in order to achieve security
and correctness guarantees. One of the significant fea-
tures of blockchain is a smart contract., which is a
set of promises and protocols agreed among parties
for the execution of the promises (Wood et al., 2014).
Participants in the network validate the execution out-
come of a smart contract. To deploy a smart contract
on the blockchain, it is compiled to get Ethereum Vir-
tual Machine (EVM) bytecode, and then a contract
creation transaction that contains bytecode is sent to
the network. Authorized participants can interact with
a contract using the application binary interface (abi).
A transaction sender needs to pay gas to invoke con-
tract functions; gas is a measure of computational ef-
fort to execute EVM operations.
For a successful IoT application on a blockchain,
a blockchain needs to handle a large number of de-
vices. However, the rate at which blockchain handles
transactions is limited. Recently, many approaches
for solving the blockchain scalability issue have been
proposed (Chauhan et al., 2018). One of the notable
approaches is a sidechain; a sidechain is a secondary
blockchain connected to the mainchain. The proto-
cols of the sidechain are designed to meet application-
specific goals. The user can offload some of the tasks
to the sidechain to achieve high transaction process-
ing speed and save high transaction fees. The inte-
gration of sidechains to the mainchain increases scal-
ability; however, the consensus algorithm is critical to
maintaining the security of the sidechain.
4 SEBDA SOLUTION
This section describes the architecture of the SEBDA
scheme.
4.1 SEBDA Overview
In this section, we introduce our SEBDA scheme at
an abstract level. First, a querier initiates the data
aggregation process by publishing a smart contract
Querier
Data
Provider
Sidechain
Mainchain
Figure 1: System overview.
mainchainAgg on a mainchain. Next, each partici-
pant register themselves by sending a transaction to
the mainchainAgg contract. Then, data providers are
divided into multiple clusters, and each cluster is as-
signed a primary cluster head and secondary cluster
heads, respectively. The primary cluster head com-
putes data aggregation and reports the result to the
mainchain. This opens a challenge period during
which secondary cluster head can challenge the sub-
mitted aggregation result. After the challenge period
expiration, the mainchainAgg contract performs in-
tercluster aggregation. A detailed explanation of the
SEBDA is described in Section 4.2.
4.2 SEBDA Protocol Details
The proposed data aggregation protocol consists of
the following ve phases: system initialization, reg-
istration, intracluster data aggregation, dispute reso-
lution, and intercluster data aggregation.
4.2.1 System Initialization
Each participant has a digital wallet, a piece of soft-
ware enabling interaction with the blockchain. A wal-
let has a master private key that can essentially gener-
ate unlimited public key pk and private key sk pairs.
A participant’s public key pk is visible to the network.
A private key sk is known only to the participant and
is used to create a digital signature of a transaction
using the Elliptic Curve Digital Signature Algorithm
(ECDSA) (Breitner and Heninger, 2019).
4.2.2 Registration
The querier initiates the aggregation process by de-
ploying a smart contract, called mainchainAgg con-
tract to the mainchain by sending a transaction. After
a successful deployment of the mainchainAgg con-
tract, it is assigned a unique address add
main
that
SEBDA: A Secure and Efficient Blockchain Based Data Aggregation Scheme
371
is used for further interactions. In addition to the
mainchainAgg contract, a querier also publishes a
contract called sidechainAgg contract on InterPlan-
etary File System (IPFS). IPFS is a peer-to-peer dis-
tributed file-sharing system. Further details of these
contracts are mentioned in subsequent sections.
Next, each data provider interested in participat-
ing registers itself by sending a transaction that in-
vokes a function regDp of the mainchainAgg con-
tract. The mainchainAgg contract mainly handles
participants’ registration, clusterization, challenges,
and intercluster aggregation functions. As a result
of successful registration, each device is assigned a
unique identifier d
i
. Participants interested in acting
as aggregators also register themselves on the main-
chain. To do so, each aggregator sends a transac-
tion that invokes a function regCh of mainchainAgg
contract. To thwart submission of incorrect aggrega-
tion results, aggregators are required to submit tokens
called deposits during registration. The deposit act
as leverage against malicious aggregators. Aggrega-
tors are also incentivized to participate and follow the
rules in the network by rewarding them.
Next, querier partitions data providers D into mu-
tually disjoint non-empty n sets whose union is D and
each set is called a cluster c
i
. Then, each cluster c
i
is assigned an aggregator called primary cluster head.
Each cluster c
i
is also assigned multiple watchdogs,
called secondary cluster heads. A secondary clus-
ter head verifies the aggregation result submitted by
primary cluster head. Both primary cluster head and
secondary cluster head are randomly selected. They
can also be chosen by aggregator election algorithm
such as LEACH protocol (Heinzelman et al., 2000;
Heinzelman et al., 2002); but this is not in the scope
of this research.
4.2.3 Intracluster Data Aggregation
In this phase, each cluster c
i
creates its sidechain. The
sidechain stores data values of data providers of a
cluster, and aggregators are responsible for manag-
ing it. After the sidechain creation, the primary clus-
ter head fetches the pointer to sidechainAgg con-
tract from the mainchain. Then, using the pointer,
the primary cluster head gets the sidechainAgg con-
tract from IPFS and publishes it on the sidechain. As
a result of successful sidechainAgg contract publi-
cation, it is assigned a unique address add
sidecontract
.
Next, the cluster head sends the address add
sidecontract
and abi of sidechainAgg contract to the data
providers of its cluster.
Each data provider d
i
in the cluster encapsulates
its data value v
i
in a transaction and sends it to the
sidechain within a specified period. The cluster ag-
gregator verifies the transaction and executes it if it
recognizes the data provider as an authenticated clus-
ter member. The state of the intracluster aggregate
changes as a result of successful transaction execu-
tion. After computing the intracluster aggregation re-
sult, the primary cluster head submits it to the main-
chain.
When the mainchain receives the submitted intra-
cluster aggregation result, it emits an event to notify
secondary cluster heads. This opens a challenge pe-
riod during which secondary cluster head of a clus-
ter can open a dispute. Each secondary cluster head
compares the submitted intracluster aggregation re-
sult with the intracluster aggregation result computed
on the sidechain; if they are not equal, the secondary
cluster head challenge it within the challenge period.
If the intracluster aggregation result has not been chal-
lenged by any secondary cluster head during the chal-
lenge period the mainchain optimally accepts it, re-
wards the primary cluster head for good behavior, and
proceeds to Section 4.2.5.
4.2.4 Dispute Resolution
This phase executes only when one of the secondary
cluster heads challenges the intracluster aggrega-
tion result submitted by primary cluster head. The
mainchainAgg contract emits an event about the in-
tracluster aggregation result submitted by challenger
secondary cluster head and asks to vote within a spec-
ified period t. If the majority of secondary cluster
head votes for the challenger aggregation result it is
accepted as a valid result; otherwise, the system ac-
cepts the intracluster aggregation result submitted by
the primary cluster head.
The dispute resolution phase ensures that the in-
correct aggregation result submitted by malicious pri-
mary cluster head is not accepted by the system
as long as the majority of aggregators are honest.
Similarly, a malicious secondary cluster head can-
not falsely win the challenge. To penalize a mali-
cious primary cluster head, a part of its deposit is
burned, meaning tokens are removed from circulation
by sending them to the wallet address that cannot be
used by anyone.
4.2.5 Intercluster Data Aggregation
Once all clusters have submitted their aggregation re-
sult and the challenge period is over, the querier initi-
ates intercluster aggregation by sending a transaction
that invokes a function computeInterclusterAgg of
mainchainAgg contract. The intercluster aggregate is
computed on the mainchain, and a network of miners
ensures transparent and correct computation of the in-
SECRYPT 2023 - 20th International Conference on Security and Cryptography
372
Algorithm 1: Challenge.
Input: cluster id c
i
,
cluster head address ch
i
,
claimed value claimedAgg
i
Modifier: is challenge period active,
is challenger valid;
1: start challenge against cluster head ch
i
with claimed
value claimedAgg
i
;
2: increment vote v against cluster head ch
i
by 1;
3: if total votes v is greater than majority of aggregators
then
4: agg
i
claimedAgg
i
5: a part of ch
i
deposit is burned
6: if ch
i
.deposit <= 0 then
7: remove cluster head ch
i
from aggregators list
8: end if
9: end if
tercluster aggregate. As a result of successful execu-
tion, an event is emitted to notify the querier about the
aggregation result.
5 SECURITY ANALYSIS
5.1 End-to-end Integrity
To show that SEBDA ensures end-to-end data in-
tegrity, we analyze it in 2 scenarios 1. incorrect intra-
cluster data aggregation 2. incorrect intercluster data
aggregation.
For the first scenario, assume there is a malicious
primary cluster head A which submits incorrect in-
tracluster aggregation result agg
i
to the mainchain.
Once the mainchain receives the result, it notifies sec-
ondary cluster heads of the cluster c
i
and opens the
challenge period. An honest secondary cluster head
challenges the malicious primary cluster head A
i
and
submits correct aggregation result claimedValue, ac-
cording to Algorithm 1. Other honest secondary clus-
ter heads belonging to cluster c
i
votes in favor of the
challenger. The intracluster aggregation result agg
i
submitted by A is replaced by claimedValue when
claimedValue receives v/sch
i
> τ votes, where v is a
total number of votes in favor of claimedValue
i
, sch
i
is a total number of secondary cluster head in the clus-
ter i and τ is threshold whose value must be greater
than 50%. The more the secondary cluster head there
are, the more challenging it becomes for a malicious
primary cluster head to compromise the aggregation
result by colluding with a significant portion of clus-
ter heads. According to the assumption, the majority
of secondary cluster head are honest; thus, intraclus-
ter aggregation result agg
i
submitted by malicious pri-
mary cluster head A is replaced by correct intracluster
aggregation result agreed by the honest majority.
Now, let’s consider the second case in which ad-
versary A tries to manipulate intercluster aggregation
result. Our scheme leverage smart contract to com-
pute intercluster aggregation result. The miners of
the main network are responsible for verifying trans-
actions and computing the new state of the smart con-
tract. Therefore, an adversary A can only manipu-
late intercluster aggregation results if its hash power
is greater than half of the total hash power, which is
hard to achieve in practice.
5.2 Authentication and Non-repudiation
All participants sign their transactions using ECDSA.
Other participants use the public key to verify the
transaction’s integrity (single-hop data integrity) and
authenticity. Single-hop integrity ensures that an ex-
ternal entity does not modify the aggregation result
but does not ensure that the aggregation result is com-
puted correctly. ECDSA has been proven existentially
unforgeable under the hardness of the Discrete Log-
arithm Problem (DLP) and collision-resistant hash
function. Furthermore, since all participants sign their
transaction using their private keys; therefore a partic-
ipant cannot deny a transaction. Thus, the proposed
scheme ensures single hop integrity, authentication
and non-repudiation properties.
5.3 False Data Injection
In a false data injection attack, an external adversary
A attempts to inject false data as data values or aggre-
gation results. It is essential to ensure false data in-
jection because the blockchain is a public ledger, and
any member of a blockchain can send a transaction
to submit data value and aggregation results. In our
scheme, each primary cluster head uses ECDSA to
signature a transaction that invokes an intracluster ag-
gregation result function subIntraclusterAgg. The
modifier of a function subIntraclusterAgg ensures
that only authenticated primary cluster head can sub-
mit intracluster aggregation results. Modifiers check
the prerequisites of a function and execute the func-
tion of a smart contract only if prerequisites are sat-
isfied. Thus the modifier of a subIntraclusterAgg
function rejects the intracluster aggregation result if
an adversary A tries to submit an intracluster aggrega-
tion result. Similarly, the sidechainAgg contract en-
sures that only authenticated data providers of a clus-
ter can submit data values. Also, the mainchainAgg
contract ensures that only members of a cluster can
challenge primary cluster head. Hence, by utilizing
ECDSA and modifiers, our scheme filters out false
data submitted by an external adversary A .
SEBDA: A Secure and Efficient Blockchain Based Data Aggregation Scheme
373
Table 1: Comparision between SEBDA and related work.
Paper Single hop
data In-
tegrity
False Data
Injection
Scalable End-to-end
data in-
tegrity
Smart contracts
(Fan et al., 2020)
(Wang et al.,
2020)
task publication
(Xie and Chen,
2021)
- data aggregation
(Loukil et al.,
2021)
group formation, aggregator selec-
tion
(Lu et al., 2021)
SEBDA data aggregation , dispute resolu-
tion
5.4 Comparision with Related Work
Comparision with other blockchain-based data aggre-
gation schemes (Fan et al., 2020; Wang et al., 2020;
Xie and Chen, 2021; Loukil et al., 2021; Lu et al.,
2021) shows that they provide single-hop data in-
tegrity and do not consider malicious aggregator; thus
fails to achieve end-to-end integrity. The main goal of
(Xie and Chen, 2021) is to protect the data confiden-
tiality in the presence of an untrusted aggregator and
leader. However, a leader responsible for calling the
aggregation function of a smart contract can submit
an incorrect or incomplete database DB where DB is
collected data of devices. Additionally, the system is
not scalable enough to handle many IoT devices.
6 EVALUATION
In this section, we compare the performance of our
solution with a Conventional Blockchain-based Data
Aggregation (CBDA) system that performs all opera-
tions on the mainchain.
6.1 Experimental Setup and Procedures
We chose Ethereum blockchain to conduct our exper-
iments, and Ganache is used to simulate the Ethereum
blockchain. To make Ganache close to the main
Ethereum network, we set the block time to 12 sec.
It is important to note that other factors, such as gas
price and the number of pending transactions, also in-
fluence the transaction processing time on Ethereum
main network. However, these factors are challenging
to incorporate into the experimental setup. The smart
contracts are written in Solidity, and the web3.js li-
brary is used to interact with the Ethereum node. The
data values of data providers are randomly generated,
and a sum aggregation query is used for performance
evaluation. We considered 3 sidechains connected to
the mainchain, and each sidechain is assigned one pri-
mary cluster head and two secondary cluster heads.
The experiments are repeated 30 times and each data
value represents the average. The experiments are
conducted on the physical machine equipped with In-
tel(R) Xeon(R) CPU E5 2660 v4, 131.7964 GB
RAM, CPU Frequency 2.00GHz, and docker con-
tainer is used for virtualization.
6.2 Results
We have evaluated the performance of our scheme
based on gas, transactional cost, and computational
cost. Table 2 shows the gas consumed by different
functions of a solidity smart contract. Ethereum sets
the minimum gas units for any transaction to 21, 000.
This mainly covers the cost of running an elliptic
curve algorithm. In addition to this, a transaction also
incurs gas that is dependent on the used opcodes in
the smart contract. It can be seen that the cluster head
registration function regCh requires the highest gas.
This is because when a node sends a transaction to
register as cluster head, the regCh function triggers
the change of Ethereum accounts state as a result of
deposit transfer. Then, it also updates the smart con-
tract’s state to reflect the newly added cluster head.
Table 2: Gas unit of different functions.
Number Function Gas units
1 regDp 95413
2 regCh 179334
3 subData 27631
4 subIntraclusterAgg 61093
5 compInterclusterAgg 76791
6 challenge 76652
Next, in Figure 2, we compare the cost of the data
aggregation process in terms of transaction fees. The
data aggregation cost is the sum of each transaction
fee executed during the aggregation process, includ-
ing the registration fee. The results are compared with
SECRYPT 2023 - 20th International Conference on Security and Cryptography
374
the CBDA system. The transaction fee is calculated
as transaction f ee = gas units gas price per unit.
The gas used by different aggregation functions can
be seen in Table 2. Gas price is the amount of ether
(ETH) per gas unit that must be paid to execute a net-
work transaction and fluctuates over time. ETH is a
digital coin used in the Ethereum network, and one
of the denominations of ETH is Gwei. At the time
of running experiments (26 Sep 2022), the gas price
was 14 Gwei in the Ethereum network; thus, we have
also used 14 Gwei as the gas price for our network. In
Figure 2, it can be observed that our proposed SEBDA
results in overall lower transaction fees as compared
to the CBDA system. This is because intracluster ag-
gregation computation is performed on the sidechain
in our SEBDA scheme. Hence, it saves transaction
fees and encourages participation of data providers in
the network.
4 ×10
7
5 ×10
7
6 ×10
7
7 ×10
7
8 ×10
7
9 ×10
7
1 ×10
8
1.1 ×10
8
1.2 ×10
8
1.3 ×10
8
1.4 ×10
8
5 10 15 20 25 30 35 40 45 50
Gas price (GWEI)
Number of data providers
CBDA
SEBDA
Figure 2: Transaction cost comparision.
Figure 3 displays the computational cost of
data submission and intercluster aggregation in our
scheme. Our SEBDA scheme takes 139 ms for
data submission and intracluster aggregation of 100
data providers. Both operations are handled on the
sidechain, improving efficiency and removing trans-
actional costs on the data provider. In Figure 4, we
illustrate the scalability of our scheme and analyze
the effect of an increasing number of data providers
on the total data aggregation computation time. The
inside smaller chart we zoomed in on the y-axis of
the original graph. In our SEBDA scheme, most of
the total time is consumed by intracluster aggregation
result submission and intercluster cluster aggregation
because these operations are done in the mainchain,
which takes 12 sec to process a single transaction as
we can see that our SEBDA scheme shows superior-
ity than the CBDA system with an increasing number
of data providers. For example, our SEBDA scheme
reduces 91.9% of the total computational cost for 50
data providers. Thus, it can be concluded that our
scheme can handle a large number of data providers
without affecting overall performance, as its compu-
tational time increases linearly with the increasing
number of data providers.
0
20
40
60
80
100
120
140
160
10 20 30 40 50 60 70 80 90 100
Time (ms)
Number of data providers
SEBDA
Figure 3: Computational cost of data submission and intra-
cluster aggregation.
7 CONCLUSION AND FUTURE
WORK
We propose a 3-tier blockchain-based data aggrega-
tion scheme that use sidechains for data submission
and intracluster aggregation. To protect the end-to-
end integrity, SEBDA utilizes challenge and dispute
resolution protocols. The reward and punishment pro-
tocols encourage participation and discourage mali-
cious behaviors. Our simulation results prove that
SEBDA improves efficiency and scalability. To the
best of our knowledge, this is the first blockchain-
based data aggregation scheme that addresses in-
tegrity and scalability requirements. In the future, we
will investigate the effect of an increasing number of
sidechains on security and find a sweet spot between
security and scalability of the blockchain-based data
aggregation scheme. Moreover, we also aim to deploy
the proposed solution on low power IoT devices and
conduct experiments in a practical setting. This will
involve creating a small-scale IoT network where IoT
devices submit data values to the sidechain.
0
200
400
600
800
1000
1200
10 20 30 40 50 60 70 80 90 100
48
48.02
48.04
48.06
48.08
48.1
48.12
48.14
20 40 60 80 100
Time (sec)
Number of data providers
CBDA
SEBDA
Figure 4: Computation cost of data aggregation.
SEBDA: A Secure and Efficient Blockchain Based Data Aggregation Scheme
375
REFERENCES
Agrawal, S. and Boneh, D. (2009). Homomorphic macs:
Mac-based integrity for network coding. In Applied
Cryptography and Network Security: 7th Interna-
tional Conference, ACNS 2009, Paris-Rocquencourt,
France, June 2-5, 2009. Proceedings 7, pages 292–
305. Springer.
Boneh, D., Lynn, B., and Shacham, H. (2001). Short sig-
natures from the weil pairing. In International confer-
ence on the theory and application of cryptology and
information security, pages 514–532. Springer.
Breitner, J. and Heninger, N. (2019). Biased nonce sense:
Lattice attacks against weak ecdsa signatures in cryp-
tocurrencies. In International Conference on Finan-
cial Cryptography and Data Security, pages 3–20.
Springer.
Catarinucci, L., De Donno, D., Mainetti, L., Palano, L., Pa-
trono, L., Stefanizzi, M. L., and Tarricone, L. (2015).
An iot-aware architecture for smart healthcare sys-
tems. IEEE internet of things journal, 2(6):515–526.
Chauhan, A., Malviya, O. P., Verma, M., and Mor, T. S.
(2018). Blockchain and scalability. In 2018 IEEE In-
ternational Conference on Software Quality, Reliabil-
ity and Security Companion (QRS-C), pages 122–128.
IEEE.
Cisco, U. (2020). Cisco annual internet report (2018–2023)
white paper. Cisco: San Jose, CA, USA.
Cui, J., Shao, L., Zhong, H., Xu, Y., and Liu, L. (2018).
Data aggregation with end-to-end confidentiality and
integrity for large-scale wireless sensor networks.
Peer-to-Peer Networking and Applications, 11:1022–
1037.
Fan, H., Liu, Y., and Zeng, Z. (2020). Decentralized
privacy-preserving data aggregation scheme for smart
grid based on blockchain. Sensors, 20(18):5282.
He, W., Liu, X., Nguyen, H., and Nahrstedt, K. (2009).
A cluster-based protocol to enforce integrity and pre-
serve privacy in data aggregation. In 2009 29th IEEE
International Conference on Distributed Computing
Systems Workshops, pages 14–19. IEEE.
He, W., Nguyen, H., Liuy, X., Nahrstedt, K., and Abdelza-
her, T. (2008). ipda: An integrity-protecting private
data aggregation scheme for wireless sensor networks.
In MILCOM 2008-2008 IEEE Military Communica-
tions Conference, pages 1–7. IEEE.
Heinzelman, W. B., Chandrakasan, A. P., and Balakrishnan,
H. (2002). An application-specific protocol architec-
ture for wireless microsensor networks. IEEE Trans-
actions on wireless communications, 1(4):660–670.
Heinzelman, W. R., Chandrakasan, A., and Balakrishnan,
H. (2000). Energy-efficient communication protocol
for wireless microsensor networks. In Proceedings of
the 33rd annual Hawaii international conference on
system sciences, pages 10–pp. IEEE.
Jeong, J. and Lee, E. (2014). Vcps: Vehicular cyber-
physical systems for smart road services. In 2014
28th International Conference on Advanced Informa-
tion Networking and Applications Workshops, pages
133–138. IEEE.
Krishnamachari, L., Estrin, D., and Wicker, S. (2002). The
impact of data aggregation in wireless sensor net-
works. In Proceedings 22nd international conference
on distributed computing systems workshops, pages
575–578. IEEE.
Leontiadis, I., Elkhiyaoui, K., Önen, M., and Molva, R.
(2015). Puda–privacy and unforgeability for data ag-
gregation. In International Conference on Cryptology
and Network Security, pages 3–18. Springer.
Li, F. and Luo, B. (2012). Preserving data integrity for
smart grid data aggregation. In 2012 IEEE Third Inter-
national Conference on Smart Grid Communications
(SmartGridComm), pages 366–371. IEEE.
Liu, X., Yu, J., Li, F., Lv, W., Wang, Y., and Cheng, X.
(2019). Data aggregation in wireless sensor networks:
from the perspective of security. IEEE Internet of
Things Journal, 7(7):6495–6513.
Loukil, F., Ghedira-Guegan, C., Boukadi, K., and Ben-
harkat, A.-N. (2021). Privacy-preserving iot data ag-
gregation based on blockchain and homomorphic en-
cryption. Sensors, 21(7):2452.
Lu, W., Ren, Z., Xu, J., and Chen, S. (2021). Edge
blockchain assisted lightweight privacy-preserving
data aggregation for smart grid. IEEE Transactions on
Network and Service Management, 18(2):1246–1259.
Nakamoto, S. (2008). Bitcoin: A peer-to-peer electronic
cash system. Decentralized Business Review, page
21260.
O’Keeffe, M. (2008). The paillier cryptosystem. Mathe-
matics Department April, 18:1–16.
Vinodha, D. and Mary Anita, E. (2021). Discrete integrity
assuring slice-based secured data aggregation scheme
for wireless sensor network (dia-ssdas). Wireless
Communications and Mobile Computing, 2021:1–17.
Wang, X., Garg, S., Lin, H., Kaddoum, G., Hu, J., and Hos-
sain, M. S. A Secure Data Aggregation Strategy in
Edge Computing and Blockchain empowered Internet
of Things. pages 1–1.
Wood, G. et al. (2014). Ethereum: A secure decentralised
generalised transaction ledger. Ethereum project yel-
low paper, 151(2014):1–32.
Xie, X. and Chen, Y.-C. Decentralized Data Aggregation: A
New Secure Framework based on Lightweight Cryp-
tographic Algorithms. In 2021 IEEE Conference on
Dependable and Secure Computing (DSC), pages 1–
2.
Xie, X. and Chen, Y.-C. (2021). Decentralized data aggre-
gation: a new secure framework based on lightweight
cryptographic algorithms. Wireless Communications
and Mobile Computing, 2021.
SECRYPT 2023 - 20th International Conference on Security and Cryptography
376