You are on page 1of 4

Utilization of the Blockchain Network in The Public

Community Health Center Medicine Supply Chain


Hartanto Kumiawan JungYeon Kim Hongtaek Ju
dep. Of computer engineering dep. Of computer engineering dep. Of computer engineering
Keimyung University Keimyung University Keimyung University
Daegu, Republic of Korea Daegu, Republic of Korea Daegu, Republic of Korea
hartantokumiawan@gmail.com 5341308@stu.kmu.ac.kr junt@kmu.ac.kr

Abstract—Distribution of medicine in the —œ‰“ŠG Œˆ“›G provides the proposed system architecture, section 4 explains
community centers has been regulated and monitored in an the detailed implementation, section 5 discusses the conducted
official distribution channel. However, counterfeit medications experiments as well as the gotten results. Finally, Section 6
still circulate in the community. Circulation of counterfeit concludes the paper.
medicines is caused by factors from businesses, consumers, and
medicine distribution supervision. The use of blockchain II. BACKGROUND
technology in supply-chain has shown new breakthroughs to
overcome these problems since it ensures immutability, A. Brief description of MSCM
transparency, and decentralization of data. In this paper, we The Medicine Supply Chain Management (MSCM) is a
proposed a new blockchain-based Medicine Supply Chain continuous multi-period process based on the patient's need
Management (MSCM) system. The proposed system has been for medicines. Therefore, the amount of drug needed in a
applied to medicine distribution channels between public certain period cannot be certain.
community health centers in Indonesia. To evaluate the
proposed system, we used JMeter as a benchmarking tool to The problem that often occurs in MSCM is the
measure the system performance in terms of transactions per distribution of medicines that are not in accordance with the
second, latency, and the utilization of resources. needs of medicine stocks in each location. It results in the
possibility that the medicine stock will be canceled at the
Keywords - blockchain, supply chain management, location where it is needed, while there is an excess of
healthcare, Hyperledger fabric, performance analysis medicine stock at the unnecessary site [6]. The entry of
medicine from unofficial suppliers is a loophole in the
I. INTRODUCTION circulation of fake medicines.
The World Health Organization (WHO) estimates that 10%
of the medicines in the world may be fake [1]. World Customs B. Combining blockchain with MSCM
Organization (WCO) also estimated that losses resulting from Blockchain is a distributed database that keeps a growing
counterfeit medicines reach $ 200 billion annually, as well as list of blocks secured from disruptions and revisions.
counterfeit pharmaceuticals accounting for more than half of Composed of blocks, holds a collection of individual
the 24,000 confiscation cases related to illegal medicines in transactions. Each block contains a timestamp and a link to the
2013 [2]. Indonesia assesses the value of counterfeit previous block [7]. This cryptographic technology offers a
medicines discovered in 2012 to be in the region of USD 61.6 way for people who do not know or do not trust each other to
million [3]. make notes about who owns what will force the agreement of
all the people concerned. It is a way to build and preserve the
Indonesia had around 9,993 units of Puskesmas (public trust [8].
community health center) in 2018, as many as 3,623 units
(36%) have inpatient services [4]. To improve services, Another perspective on extracting potential blockchain as
currently, several Puskesmas in cities and districts have begun a case of using "original" blockchain is source tracking. It
to use the health system that have MSCM features in seems that tracking the origins along the MSCM can be one
accordance with the regulations of the Republic of Indonesia's of the blockchain applications.
Ministry of Health [5]. The existing health system in the
Puskesmas is diverse and still has weaknesses, especially in III. SYSTEM DESIGN
the MSCM feature, for example, the low accuracy of data and A. The Schema of the Medicine Supply Chain
manipulation of medicine inventory data.
Figure 1 represents the block chain that manages the
This study aims to use the blockchain technology in supply chain and stores data related to medicines,
pharmaceutical supply chains with true and reliable data for manufacturer, distributors, health office, public hospitals,
Puskesmas. First, it aims to propose an integrated system for purchase orders, and commodity transfer data, among the
conducting and recording medicine supply transactions on the trusted partners who apply consensus to maintain the
blockchain network with MSCM. Data is distributed to all the consistency of the distributed ledger.
participants in the network and it is measured according to the
time needed for fully propagated. Furthermore, we discuss the
issues related to access rights in the blockchain system to
maintain data security. Finally, the efficiency and reliability of
the system in the process of retrieving data to track historical
information in a chain are presented.
The paper is organized as follows: Section 2 introduces
the background of MSCM and blockchain with a discussion
of the application of the blockchain in the MSCM, section 3

&RS\ULJKW.,&6  $31206


Authorized licensed use limited to: University of Gothenburg. Downloaded on August 31,2021 at 15:36:00 UTC from IEEE Xplore. Restrictions apply.
Fig. 3. PO and commodity transfer activity diagram.

IV. IMPLEMENTATION.
Fig. 1. Medicine distribution supply chain management scenarios in
the proposed blockchain system. A. . Development Environment
Table 1 describes the software and hardware used during
B. The Proposed System Architecture of the Medicine the implement of the blockchain-based MSCM system.
Supply Chain
Necessary data is stored on a distributed ledger in blocks, TABLE I. PROPOSED BLOCKCHAIN BACK-END SERVER MEDICINE
were each block consists of several transactions with SUPPLY CHAIN

encrypted hashes of previous blocks forming a chain of linked Component Description


blocks. Transaction proposals are submitted by end-users (i.e. IDE-tools Visual Studio Code v1.39.1
health offices, public hospitals, factories, distributors, retailers) Hyperledger-Fabric version v1.2
by calling back-end providers through the blockchain network Node-version v10.15.3
shown in Figure 2. CLI Tool Composer REST server v0.19
Docker-Compose Version 1.13.0
Python v.2.7.15+
Docker-Engine version Version 19.03.12-rc1
CPU Intel Core i5-7500T CPU @2.70GHz
Operating Systems Ubuntu Linux 18.04 LTS
Memory 8GB

B. Network Requirements and Topology


The system requires a stable Internet connection with at
least 512kb/s, to create any blockchain transactions. The basic
elements of the blockchain network are peers, which write
ledgers and smart contracts to carry out operations.
Fig. 2. System architecture of the proposed medicine supply chain
Hyperledger Fabric also supports the creation of multiple
management.
channels in the blockchain. Channels are blockchain safe
paths used to separate and differentiate data for a group of
The data access control can be defined by many parties,
peers.
depending on different scenarios. For instance, after initiating
the order, the public hospital needs to be able to define who C. The Chaincode of the Medicine Supply Chain
has access to the order results. The data viewer should be able The proposed system uses the Hyperledger Composer
to define access to data if the data need to be transferred to v0.19 for the implementation of the blockchain network. The
another individual. The service provider should also have used blockchain network elements were created as composer
control over the data because it provides the service. files. The participants, assets, and transactions proposed in the
C. The Medicine Supply Chain Transaction Process solution structure were created as a model file (.cto) using the
Hyperledger Composer Modeling Language. The transaction
The activity diagram shows the flow of information
logic was created as a logic file (.js) using Javascript. The
between the main entities in the system. Medicine supply
access control rules were created as an access control file (.acl)
chain transactions are divided between peers, and transactions
using the Hyperledger Composer Access Control Language.
are carried out by peers in the world state. Up to the order
The queries were created as the language used to return data
transaction level, several transactions can be processed
about the blockchain world-state (.qry) [9].
simultaneously by each node. The role of each component in
the network in the drug supply chain is displayed and the The initiatePurchaseOrder() function (Algorithm 1) is
ordering transaction process is shown in Figure 3 where the invoked with InitiatePO. This function starts by calling the
Health Service initiates an order. getFactory() function to search and find network servers and
suppliers and then calls the getCurrentParticipant () function.
The order will ask for the network server with the ID from
InitiatePO, then enter the order item and provide the order
status if conditions are met to return to the assetRegistry to add
the order.


Authorized licensed use limited to: University of Gothenburg. Downloaded on August 31,2021 at 15:36:00 UTC from IEEE Xplore. Restrictions apply.
Algorithm 1. Insert Purchase Order Information one client condition provides the baseline performance for the
1: Function initiatePurchaseOrder(InitiatePO) solution.
2: factory <- getFactory()
3: NS <- 'org.sipchain.network' A. . Experiment Method
4: me <- getCurrentParticipant()
5: order <- factory.newResource(NS, 'PO', InitiatePO.orderId)
The test plan was set up in the JMeter GUI mode and the
6: order.itemList <- InitiatePO.itemList test was run in the no- GUI mode, as suggested by the JMeter
7: if (InitiatePO.orderTotalPrice) then Manual. First, the server was set to run on the NGROK Cloud
8: order.orderTotalPrice <- InitiatePO.orderTotalPrice service, then the 100 sample participants and sample initiate
9: order.orderStatus <- 'INITIATED' PO were created. Afterwards, one client condition was
10: order.orderer <- me administered for the two Get methods and two Post methods.
11: order.vendor <- InitiatePO.vendor
12: getAssetRegistry(order.getFullyQualifiedType()).then(function Then the 25, 50 and 100-client condition was administered for
(assetRegistry) the two Get and two Post methods, each of the clients targeted
13: return assetRegistry.add(order) one unique resource under each request. each of the clients
targeted one unique resource under each request. For all the
D. Implementation Result measurements, the ramp-up period was set to 1s.
When query a medicine data, the REST composer of
Hyperledger server responds to the query in JSON format as In addition, to estimate the difference between the Get
shown in Figure 4. request and the Post request, the measurements under the three
levels of client numbers can be taken as three different
measurements to calculate the average, standard deviation
(SD), 95% confidence intervals (CI), and the 99% CI for the
difference.
B. Experiment Results
Overall, for the Get and Post methods, data were averaged
over each of the two requests for each number of client
conditions. It provides a more accurate measurement of the
overall performance compared with a measurement of a single
request. The overall performance is shown in Figure 5. First,
The relatively slower Post request compared with the Get
request can be seen across the whole range of the number of
clients tested. This confirms the prediction that Get requests
Fig. 4. Interface of REST Composer server for fetching medicine are faster than Post requests. Writing and updating the ledger
records. status on the blockchain is the first reason that new
information is a Post request, which requires extra time.
V. EXPERIMENT Second, if the number of clients increases, so will the latency.
The experiment uses two independent variables, the Thus, the Get and Post requests confirm the second prediction,
request type and the number of clients. First, there are two which estimates that as the total number of clients increases,
major HTTP methods related to the current application, and performance will decrease (i.e., increase in response latency)
they are Get and Post. Different methods may have different keeping in mind that each additional number of clients
performances. More specifically, the Get method should be requires a certain amount of time and resources.
faster because it is acquiring information from the blockchain,
which does not change the world state of the common ledger.
The Post method should be slower because it requires that
information be written to the blockchain, which will change
the world state of the common ledger. Therefore, the type of
method was selected as an independent variable. Second, the
number of clients that are making requests can affect the
performance. When the number of clients increases, then the
performance may decrease. As a result, the number of clients
was selected as another independent variable. Note that the
number of nodes was not taken as a directly manipulated
variable for several reasons. Consequently, the time acquired
can be used as the estimation of the time to write information Fig. 5. The overall performance by the number of clients and request
to the blockchain for an additional node. This is because each type
node functions similarly. For the Get request, because data can
be acquired from one node, there should not be a significant Figure 6 describes the average latency of the proposed
difference between different numbers of nodes. system, min, max, and percentile. By querying transactions
As mentioned above, the used HTTP methods are Get and with three different groups, this test aims to measure the
Post. The number of clients was divided into three levels: 25, blockchain network latency as proposed. As the results are
50, and 100 clients. The three levels here help to determine shown in the figure, if the number of clients increases, the
whether the performance drops linearly or not when the system latency will increase.
number of clients has increased. The performance under the


Authorized licensed use limited to: University of Gothenburg. Downloaded on August 31,2021 at 15:36:00 UTC from IEEE Xplore. Restrictions apply.
VI. CONCLUSION
In this paper, we have introduced blockchain technology
in MSCM at Puskemas. Specifically, we studied the MSCM
requirements when adopting blockchain technology and also
show the technical challenges when designing blockchain for
drug supply chains. A Case study is presented to address the
problem and introduce the medicine safety tracking system
that we are proposing. We implemented this system based on
the permissioned supply chain for supply chain scenario
medicine. Performance of various types of requests and
numbers of clients has been evaluated. The results show that
Get requests are faster than Post requests and latency increases
Fig. 6. The latency in a query transaction (Get Request) as the number of customers do. The Get request latency is, on
average, 21 seconds (SD = 22 seconds, 95% CI [-4 seconds to
In addition, the effect of the number of nodes was 46 seconds]) faster than Post requests that reflect the time for
estimated. The overall difference between the Post and Get the blockchain system to write and change ledger status
requests is shown in Figure 7. The figure was acquired by information on one node colleague. The proposed solution
subtracting the overall Get request latency from the overall provides a new way of managing data in the medical supply
Post request latency. The overall average difference across the chain to achieve reasonable efficiency.
whole number of clients is 21 seconds, with an SD of 22
seconds. The 95% CI (±1.96SE) calculated is from -4 seconds ACKNOWLEDGMENT
to 46 seconds. The 99% CI (±2.58SE) calculated is from -12 This research was supported by Basic Science Research
seconds to 54 seconds. This is an estimate for the time taken Program through the National Research Foundation of Korea
by the blockchain to write and change ledger status (NRF) funded by the Ministry of Education (NRF-
information on one peer node. Thus, if an additional node is 2018R1D1A1B07050380), and Institute for Information &
added, it functions similarly as this node, given that different communications Technology Promotion (IITP) grant funded
nodes are running in parallel and independently. As a result, by the Korea government (MSIP) (No.2018-0-00539,
there will not be a significant increase in latency. Development of Blockchain Transaction Monitoring and
Analysis Technology).
REFERENCES
[1] Gibson, L.; Drug regulators study global treaty to tackle counterfeit
drugs. BMJ (Clinical research ed.).March 2004; 328(03):486.
[2] World Customs Organization. The illicit trade report 2014. report.
Hongkong: World Customs Organizations, Department of education.
2015.
[3] Mahadi, T. Counterfeit drugs control 20% of Indonesia’s drug
circulation. Available online: https://industri.kontan.co.id/news/obat-
palsu-kuasai-20-peredaran-obat-indonesia (accessed on 1 January
2019).
[4] Kusnandar, V.B. Berapa Jumlah Puskesmas di Indonesia?. Available
Fig. 7. The overall difference between the Post and Get requests online: https://databoks.katadata.co.id/datapublish/2020/01/12/berapa-
jumlah-puskesmas-di-indonesia (accessed on 1 April 2020).
In some of these experiments several steps have been [5] Pusdatin. Sosialisasi Sistem Informasi Puskesmas. Available online:
implemented to test and analyze transaction response times, https://pusdatin.kemkes.go.id/article/view/19110500001/sosialisasi-
latency, and resource usage of the proposed system sistem-informasi-puskesmas.html (accessed on 1 April 2020).
performance. The results of these experiments have shown [6] Wahana, A; Maylawati, D; Irfan, M; Effendy, H. Supply chain
that using the blockchain technology system for supply chain management using fp-growth algorithm for medicine distribution.
can improve throughput performance while minimizing Journal of Physics: Conference Series. 978. 012018. 10.1088/1742-
6596/978/1/012018. 2018.
latency by using fewer resources. The possibilities for future
[7] Nakamoto, S. Bitcoin: a peer‐to‐peer electronic cash system. Available
testing include increasing the network size and testing its online: https://bitcoin.org/bitcoin.pdf (accessed on 10 March 2020).
efficiency and feasibility show in Figure 8. [8] The Economist Staff. The great chain of being sure about things. The
Economist. Available online:
https://www.economist.com/briefing/2015/10/31/the-great-chain-of-
being-sure-about-things (accessed on 27 February 2020).
[9] Dhillon, V; Metcalf, D; Hooper, M. Blockchain Enabled Applications:
Understand the Blockchain Ecosystem and How to Make it Work for
You. 1st ed. Springer, 2017.

Fig. 8. Latency and throughput measurement


Authorized licensed use limited to: University of Gothenburg. Downloaded on August 31,2021 at 15:36:00 UTC from IEEE Xplore. Restrictions apply.

You might also like