Blockchain based Secured Virtual Machine Image Monitor
Srijita Basu
1
a
, Sandip Karmakar
2 b
and Debasish Bera
1 c
1
Department of Computer Science and Engineering, Indian Institute of Information Technology, Kalyani, India
2
Department of Computer Science and Engineering, National Institute of Technology, Durgapur, India
Keywords:
Virtual Machine Image, Blockchai, Smart Contract, Auditability, Cloud.
Abstract:
Blockchain technology supports data immutability. Whereas, smart contracts are piece of self-executable
codes running inside the blockchain network, responsible for the transformation or state change of these
data. Furthermore, Cloud Computing is used in the application of data storage and usage. Several business
enterprises use cloud for hosting their applications and data with a minimized effort, cost and hurdles of
maintenance. However, ensuring security of client data and proper management of the Service Provider’s
infrastructure remains a crucial issue. In this article, an Ethereum based blockchain network has been proposed
that monitors and assures the safety of the Virtual Machine Images (VMI) stored at the Cloud Service Provider
(CSP) end. The proposed scheme tends to design a dedicated Smart Contract which handles each and every
function, starting from request of a VMI by the Cloud Service Consumer (CSC) to the usage of the same by
the later. The use of blockchain technology ensures that no single admin/third party can control/modify the
system. This prevents unwanted modification of the VMIs by an intruder and guarantees the efficiency of the
scheme to be higher than any other methodology designed for the same purpose till date.
1 INTRODUCTION
Cloud based services have become very popular
among small and medium scale enterprises. It re-
duces not only the hardware and software mainte-
nance costs, but also facilitates the CSC with vari-
ous add-on features like on-demand scaling, pay per
use pricing model, flexible and easy deployment op-
tions, etc. Cloud computing has drastically changed
the sphere of renting services, by providing rents on
softwares, operating environment, and even hardware
resources. A substantial apprehension comes with the
cloud based service usage in terms of security for both
the CSC’s data as well as CSP’s infrastructure.
The various security concerns (Hashizume et al.,
2013) include data locality, data segregation, Virtual
Machine (VM) life cycle management, and public
VM Image (VMI) repository regulation etc. In spite
of the fact that several cloud specific schemes (Wan
and Jiang, 2010) have been devised for each of these
issues, there are still some areas which lacks signif-
icant work. One such vital security issue is lack of
management and tracking of the Virtual Machine Im-
a
https://orcid.org/0000-0002-6835-947X
b
https://orcid.org/0000-0002-8150-3026
c
https://orcid.org/0000-0001-8888-6042
ages (VMI), which results in the sustainability of var-
ious malicious VMIs in the CSP’s data-centre.
Blockchain technology has been projected as one
of the most secured solutions for data storage and
monitoring. It has quite often been coined as a dis-
tributed and decentralized public ledger. Back in the
year 2008, Satoshi Nakamoto had used Blockchain
technology for constructing a crypto-currency plat-
form known as Bitcoin (Nakamoto, 2008). Later in
the year 2014, it was realized that blockchain could
be used to permanently record any kind of transac-
tion/change without a third party intervention. This
was followed by the 2nd generation of Blockchain-
Ethereum. Ethereum (Wood, 2014) introduced the
tech-world to Smart Contracts (Macrinici et al.,
2018). This new generation blockchain technology
helps not only digital transactions but any service, as-
set, bond etc. can be exchanged between peers, de-
pending on conditions that were depicted in the smart
contracts. The blockchain architecture (Halpin. and
Piekarska, 2017), with the help of mining algorithms
(Wang et al., 2019), clearly portrays the difficulty for
an intruder would face if he/she wishes to change
the entries of the block. Thus, it is quite evident
that blockchain with its smart contracts can be widely
used to ensure secured data/service preservation.
432
Basu, S., Karmakar, S. and Bera, D.
Blockchain based Secured Virtual Machine Image Monitor.
DOI: 10.5220/0010228804320439
In Proceedings of the 7th International Conference on Information Systems Security and Privacy (ICISSP 2021), pages 432-439
ISBN: 978-989-758-491-6
Copyright
c
2021 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
In this paper an effort has been made to design
an Ethereum (Blockchain) based Smart Contract that
would ensure VMI security and tracking. The use
of smart contract based blockchain technology in this
system ensures that no malicious VMI exists at the
CSP end and additionally tracks the entire life cycle
of a VM (Schwarzkopf, 2015) by providing a series of
authentic audit data (timestamp, violation action, etc)
which helps to identify the intended activity. The rest
of the paper is organized as follows. In Section II, a
survey of related work is presented. Section III dis-
cusses the architecture of the proposed system. Sec-
tion IV details the proposed methodology for VMI
management and tracking. In Section V result analy-
sis for the proposed scheme has been performed. Fi-
nally, Section VI concludes the article.
2 RELATED WORK
This section presents a brief survey of different
schemes on VM image tracking and security in a
cloud environment. Virtual machine images require
full-proof integrity schemes, as they initialize a VM
from scratch including its configuration as well as se-
curity parameters or policies which if changed can
pose threat to the client applications or data.
In (Wei et al., 2009), an image management
scheme was proposed for secured cloud environ-
ment. The various sub-operations /sub-modules in
this scheme include access control framework, Image
filters, provenance tracking mechanism and a set of
repository maintenance functions. The image prove-
nance tracking system was used by the CSP to ad-
dress two major issues viz. tracing the introduction
of illegal or malicious content in a VMI and delega-
tion of patching information. The proposed scheme
achieved 4.9 times speedup and addressed VMI secu-
rity aspects from both the image publisher as well as
retriever sides. Though the scheme offered a proper
tracking mechanism, it did not alleviate the risk of
uploading a VMI with malicious content.
In (Kazim et al., 2013), Kazim et. al proposed
an Encrypted Virtual Disk Images in Cloud (EVDIC)
based on OpenStack Platform. The system ensured
the protection of the stored virtual machine disk im-
ages at the Cloud Service Provider’s end. The de-
sign consisted of a Virtual Machine Image encryp-
tion module, Virtual Machine Image disk decryption
module, and Key management Module. The scheme
aimed at protecting the disk images from getting com-
promised by some malicious internal user or adminis-
trator by unauthorized access. Though the work tends
to provide security to the Virtual Machine Images
from the malicious CSP users, it is unable to protect
the same from the external intruders/ malicious Cloud
Service Consumers.
A security framework that tried to protect the Vir-
tual Machine Images (VMI) in a cloud platform was
devised by Hussein, Alenezi, Wills and Walters in
(Hussein et al., 2016). The paper formulated a re-
search path following which VMI security could be
strengthened. A model has was suggested where the
industry based security controls have been taken in
parallel with the academic security controls. At the
later stage, the redundant security controls have been
eliminated by semantically comparing each of the in-
dustry and academic controls. Finally, a combined
security control model has been produced that is ex-
pected to meet the VMI security requirements in a
better way.
It is evident from the above discussion that the
existing VM/VMI management models are either in-
complete or depend on some central or third party au-
thority/CSP admin. Most of them have certain over-
heads and are unable to present a fool-proof solution
to CSP and/or CSC. Therefore, there is a need to de-
sign a comprehensive model constructed on the core
of blockchain network, that can not only identify the
stakeholders, processes and properties of the cloud
system involved, but also provide the required track-
ing and management to ensure a secured VMI repos-
itory.
3 SYSTEM MODEL
In this section, the basic system design is outlined
with an overall insight into the entities of the pro-
posed model. The well-established definition of cloud
(Buyya et al., 2009) portrays it as a distributed system
consisting of a collection of interconnected and vir-
tualized computers that are dynamically provisioned
and presented as one or more unified computing re-
source(s) (Buyya et al., 2009). The problem that is be-
ing dealt with in this article requires the system model
to be designed accordingly.
The proposed system has been modelled to con-
tain the following components. It might be noted that
this composition is done according to the proposed
work. However, there may be some additional enti-
ties which have been ignored here due to its lack of
relevance with the present work. Fig. 1 describes the
architectural design of the proposed model, describ-
ing every component as follows.
Data-centre: A data-centre is a physical repos-
itory meant for gathering cloud computing re-
sources and other existing components (Sahli
Blockchain based Secured Virtual Machine Image Monitor
433
et al., 2014)
Hosts/Server: These are the real/physical infras-
tructures hosted inside a data-centre, used to serve
the computation and execution requests coming
from the client end (Sahli et al., 2014).
Virtual Machine (VM): It is the virtual abstraction
of the underlying infrastructure(Sahli et al., 2014)
that exists inside a host. Multiple VMs can be
instantiated as well as powered off or even killed
i.e. deleted on-demand inside single host as per
the user requests.
Virtual Machine Image (VMI): These are the var-
ious flavours of images (customized Operating
System images, sometimes pre-loaded with cer-
tain applications) that are used to boot a VM on
requirement.
Cloud Storage: Different kinds of essential data
are stored in logical pools. The physical storage is
generally sketched in two ways (Abbadi, 2011). i)
It is distributed across different dedicated storage
servers that may be located in a replicated form in
geographically distant data-centres.
ii) Again in some instances of lesser storage re-
quirements, the storage server may be a part of
computation server itself (the one that hosts the
VM for running various applications).
Block: A Block (Conti et al., 2018) is developed
from VM Request and Response documents. A
single block spans an entire VM request from
its initiation to completion (details will be dis-
cussed in the upcoming sections). In addition to
the main block there exists side block for every
main block which contains the smart contract re-
ports (explained in later section).
Blockchain: As already mentioned, each block
represents each VM access request. A sequence
of such blocks forming a chain represents a se-
ries of requests from a particular user on different
VMs.
Cloud Network: This can be physical as well as
virtual network. Since, the blockchain technology
is being used in this model the network as a whole
represents a blockchain P2P(peer-to-peer) model
where each cloud host/server and consumer con-
tains/shares the same copy of blockchain and
communicates accordingly as and when required
at the time of each blockchain consensus (Wang
et al., 2019).
Smart Contract: Smart Contracts (Macrinici
et al., 2018) add enhanced feature to normal
blockchain network. They consist of a set of self-
executable code snippets embedded as a part of
the blockchain itself. In this model, how the Smart
Contracts control the entire process of VMI re-
quest and allocation, followed by the post alloca-
tion actions has been portrayed.
Figure 1: System Mode.
4 WORKING OF THE
BLOCKCHAIN BASED VMI
MONITORING SYSTEM
In this section, VMI Monitoring scheme is presented
with its detailed working methodology. The specific
issues that can be addressed by this scheme have been
identified as follows.
4.1 Methodology
Authentication Phase: Here the user registers it-
self with the CSP and gets his/her name enlisted
as the CSC. In this Registration process, the user
needs to supply some of the basic information like
Name, Email, Contact, OrganizationName, Orga-
nization location and some additional information
if required, such as period of registration, spe-
cific purpose/project for which the user is regis-
tering etc. After this, the CSC logs in and sends
its request in the form of specific VM instance re-
quirement. E.g. the request contains the particular
Storage, RAM, CPU cycle, OS, etc. requirements.
In addition to that, the CSC might also add some
specific security requirements with respect to this
VM request. These requirements mainly depends
on the purpose for which the VM was requested.
ICISSP 2021 - 7th International Conference on Information Systems Security and Privacy
434
The kind of data to be deployed or applications
to be executed on the VM directly influences this
security requirements. After framing the request
packet, it is send to the CSP.
If the system notices that the user is not a regis-
tered CSC then the login process fails. Addition-
ally the system can also revert back/deny any ab-
normal request packet by analysing the requested
parameters (Storage, CPU cycles). The request
packet passing through all these process is for-
warded to the next phase after appending the re-
ceived request timestamp and the CSC ID with it.
VM Allocation Phase: The request is pro-
cessed for the first time in this phase. The re-
quested parameters are examined and accordingly
a VMI is selected from the public VMI repos-
itory. The VMI should be such that it fulfils
all the CSC requirements along with any spe-
cific security requirement the CSC has included.
For executing this allocation, the existing “vir-
tual
machine basic mapping” (explained in the
later section) is checked to find the possible VMIs
with permissible hardware and software config-
urations. A list of VMIs is returned, out of
which one is selected, maintaining the load bal-
ancing (Ferris, 2014) criteria of the CSP. Af-
ter selecting the VMI, system creates a VMI re-
quest document. This document includes the
hashed timestamp of the received request, hash
of the VM ID (VM ID of any VM can be re-
trieved from the System Database), hash of the
host ID that deploys the concerned VM, hash of
the CSC ID initiating the request, and finally the
hashed data containing the purpose of request and
any specific security requirements of the CSC if
any. The necessary changes according to this
request document is made in the existing “vir-
tual machine allocation” mapping (explained in
the later section). The timestamp, CSC ID, VM
ID, Host ID and the additional security data if
any, are inserted into this mapping. The basic
details for the particular VMI is also updated in
the “virtua
machine basic” mapping, as the free
space, RAM, CPU cycles and any other resource
would now be changed with the new allocation
being done. Once all the changes have been made,
the time stamp for the same is also noted as the re-
sponse time stamp.
VM Response Phase: In this step, the system con-
structs the VMI response document. This docu-
ment contains the hash of the timestamp of the
received response i.e. the timestamp when the al-
location was reflected in the blockchain, the hash
of the allocated VM login details (VM IP address
and password) and the hashed VM metadata. The
system logs the VM Request and Response docu-
ments in the blockchain (for the particular user) as
a new block. The block is appended in a chrono-
logical way, i.e. this block appears after the last
block which had been formed with respect to the
previous request from the same user. The change
of state of each VMI can now be easily tracked
from the blockchain network.
VM Action Triggering Phase: The system now ex-
ecutes necessary function using which the CSC
can log in to the VMI, allocated to him/her us-
ing the valid VM IP and password. On success-
ful login, the system logs every action as an event
into the blockchain. For every action it consults
the Blockchain Database to return the policy list
mapped against this particular VMI. After fetch-
ing the list, it is compared with the most recent
CSC action. If this action is found to be a valid
one then no step is taken otherwise for every ille-
gitimate action the same is logged as an event with
the particular time stamp and CSC ID in the side
block. This phenomenon ensures auditability and
a rigorous monitoring of the VMI. Moreover, the
CSC is forcefully logged out by ending its session
and depending upon the sensitivity of the VMI,
the password for the same might also be modified
so that the CSC is unable to login again. In case of
lower VMI sensitivity the CSC is given a chance
to re-login in the VMI with the same credentials,
but this time the particular object for which access
was denied during the previous check is locked.
The above scenario mainly concentrates on the
immediate action taken against any illegitimate
access request. Along with this, the normal mod-
ifications conducted on the VM should also be
noted. E.g. the user might install a new soft-
ware in the allocated VM or he/she might update
the version of some pre-installed software. Every
such action must be logged into the blockchain.
At the same time the overhead of this event log-
ging activities in the blockchain should be kept
as minimal as possible (as each transaction in-
volves consumption of Ether Gas (Wood, 2014)).
Keeping this in mind, the proposed system runs a
periodic blockchain write, which logs all the le-
gitimate VM modification events for that partic-
ular period into the blockchain and modifies its
policy list accordingly for that VM. This mini-
mizes the overhead of saving every change imme-
diately as it occurs in a particular VM and reflect-
ing the same in the policy list. Though the peri-
odic updates save transaction charges that would
have been incurred if continuous logging would
Blockchain based Secured Virtual Machine Image Monitor
435
have taken place, it induces another risk. Imag-
ine a scenario where User A is working with VM
X and some valid modifications have been made
by him/her on VM X. The periodic update is yet
to occur and in the meantime another user, User
B is also allocated VM X. It should be noted that
the actual version of VM X is different from what
the CSP checks before allocating the same to User
B. Moreover, the policy list with respect to VM X
has not been updated. This might be counted as a
vulnerability of the system. There might be cer-
tain unlogged changes in VM X and the Policy
list that needs to addressed against this new allo-
cation. A solution to this issue has been devised
where the system in the VM Allocation Phase
checks the following set of data viz. i) Whether
the allocated VM is already shared by some other
user at present. If the answer to this query is yes
then ii) Check the last update timestamp in the
blockchain, for the concerned VM (VM X). If this
timestamp is found to be right after the allocation
request was made, then no further action needs to
be taken and VM X is allocated to User B. But
if this timestamp is found to be before the allo-
cation request then the allocation is halted for the
time being. A forceful blockchain write is made
against the action User A-> VM X. After this the
new allocation of User B-> VM X is made.
Thus for the entire procedure of sending a request
by the CSC to fulfilment of the same, it has been
observed how the blockchain network and the em-
bedded smart contract not only helps to monitor the
VMIs across its lifecycle but also prevents any mali-
cious user to bring unintended changes to the VMI.
On one hand it secures the CSP infrastructure by as-
suring that the VMIs stored in the CSP public repos-
itory are safe. On the other hand any legitimate CSC
who is allocated a VM from this public repository can
now be sure enough to have received an uninfected
VM. Moreover, any unnatural change in a particular
VMI is auditable from the information present in the
side blocks. This information can be utilized by the
CSP, even at a later stage to decide directly whether
to accept a particular CSC request or to reject it as this
CSC might have been involved in some improper ac-
tions pertaining to any previous request which could
be easily tracked from the side blocks. Therefore, it
is evident that this scheme is beneficial from both the
CSC as well as the CSP end.
4.2 Smart Contracts
The structure and role of the smart contract has been
discussed elaborately in this sections. As described
Table 1: Smart Contract Variables.
Name and Type of
Variable
Purpose and Nature
address public
csc
ID
Stores the address of
CSC at the time of SLA
formation(S)
uint public vm ID VM/VMI identifier(S)
mapping(uint-
> CSC
request)
publicvir-
tual machine request;
Stores the various re-
quest parameter against
each CSC Identifier(S)
mapping(uint->
VMI
UI) publicvir-
tua machine allocation;
Stores the allocation
mapping of the VM
against each user(S)
uint public event ts Stores the timestamp of
the most current event
encountered by the CSC
(L)
String public ac-
tion
name
Stores the name of the
most current event en-
countered by the CSC
(L)
mapping(uint-
> VMI policy)
publicvir-
tual
machine policy;
Stores the Policy details
against each VM(S)
String public VM IP; Stores the VM IP ad-
dress (L)
String
VM
password;
Stores the VM IP pass-
word (L)
enum VM state {
Idle;
Busy;
Dead; }
Stores the different
states of a Virtual Ma-
chinee(State Variable)
earlier, a smart contract serves as the main driv-
ing agent behind a successful and secure VM allo-
cation and monitoring process. The variables used
in this smart contract have been depicted in Table
I along with their name, purpose and nature (State
(S)/Local(L) (Anonymous, 2019)). Some of the vi-
tal functions and events have also been elaborated as
follows:
Event Login: It logs/stores the user id and the
Time stamp at which a valid CSC logs in to his/her
allocated VM.
Function New Event: The function checks for the
VM and user sensitivity. It also checks whether
ICISSP 2021 - 7th International Conference on Information Systems Security and Privacy
436
the VM state is idle and the sender of this function
is a valid CSP. If all the conditions are true then
it stores the action that the user performs on the
VM as an event in a periodic manner as mentioned
in theVM Action triggering Phase in the previous
subsection. The event contains the action name,
timestamp, User ID and VM ID
Func get
VMI policy: It returns the policy
id and the read deny policy containing the read
deny policies for the particular VM. The same
would have been write
deny policy and exe-
cute deny policy while checking the write and ex-
ecute deny policy list. So for each VM 3 deny
policy lists are maintained.
Func Check policy: The function matches the
deny
policy list returned from the previous func-
tion with EName (action/event name) that the
CSS had just performed. If no match occurs, then
action can be allowed and the action is unblocked
with respect to this user where as if a match oc-
curs the function constructs a message “Uid + de-
nied access to + EName”. Now depending upon
the sensitivity of the VM two set of actions might
be possible, i) Event “Deny Access” takes place
that logs the above message along the particular
VM ID and the CSC is forcefully logged out of
the system, and the login password for the same
is changed so that the CSC might not re-login, or
ii) Event “Deny Access” takes place that logs the
above message along the particular VM ID and
the CSC is forcefully logged out of the system. In
the 2nd case the VM sensitivity is low and there-
fore the user is not permanently unallocated from
the VM. He/she is just temporarily logged out and
prevented from performing this particular action.
The functions associated with the Smart Contract that
mainly controls the VM Action Triggering Phase have
been separately depicted in Fig. 2. It acts as the core
control of the Smart Contract.
5 RESULT ANALYSIS
Here an Ethereum (Wood, 2014) based Smart Con-
tract with Solidity (Dannen, 2017) (Version 0.4.4)
over a Remix (Yann300, 2020) platform has been
used to design the intended system. In the private
blockchain setup the average Gas price has been taken
as 0.00000002 Ether, and 1 Ether =391.933 USD was
observed at the time of experiment. Table III shows
the cost associated with the functions that are mainly
involved in ensuring the VMI security as well as mon-
itoring.
Figure 2: Flow Chart for VMI Monitor Smart Contract.
Table 2: Associated Cost.
Function Gas
Used
Cost
(Ether)
USD
Event Login 13256 0.00079536 .31
Function New
Event
41567 0.00207835 0.82
Func get VMI pol-
icy
12754 0.00038262 0.15
Func Check policy 50153 0.00250765 0.98
The effect of the number of VMIs at the CSP end
on the Gas cost has been plotted in Fig 3. It has been
observed that the Gas cost increases linearly with the
number of VMIs deployed in the system. Another
observation has also been made in Fig 4., that shows
a linear relationship again, between the total no. of
policy statements in the Blockchain database and the
total Gas cost incurred.
In (Kazim et al., 2013), managing security of Vir-
tual Machine Images in a cloud environment has been
observed that there are mainly three facets of Virtual
Blockchain based Secured Virtual Machine Image Monitor
437
Figure 3: No. of VMIs v/s Gas Cost.
Figure 4: No. of Policy Statement v/s Gas Cost.
Machine Image that has been taken care of. The sys-
tem proposes and combines three different systems
each for access control, image filtering, and prove-
nance tracking. In a real life scenario, the minimal
size of the VMI is approximately 10 GB. In gen-
eral, on an average, more than 5000 files run on each
VMI (when deployed as VM). A CSP (private) with
a medium sized data centre running at least 500 such
VM instances (Marr and Kowalski, 2014). Therefore,
running and maintaining the above model at the CSP
end would result in a high percentage of over-head.
The access control policies have inevitable dependen-
cies on their owner as proposed by the system. This
creates a bottle neck and makes it vulnerable at the ad-
ministrator/owner level. Once the central authority is
compromised the entire system loses and its viability.
Finally, the provenance tracking system is designed
to track the derivation history of 500 VMIs separately.
Each new VMI can experience a huge number of state
transitions in its entire lifespan. Any centralized sys-
tem, maintaining this record, would encounter a high
number of system entries triggered by each such tran-
sition.
On a contrary, the Blockchain based VMI Mon-
itor, proposed in this article, assures that no change
in any VMI occurs without the consent of the CSP or
all the concerned hosts that are part of the blockchain
network. The Smart Contract, as already seen, is a set
of auto executable codes that run when some specific
criteria/conditions are met. Here, the blockchain net-
work with its embedded smart contract, i) Maintains
the access control policies for the VMIs, ii) Takes the
post violation decision depending on its in-built code,
and iii) Monitors or tracks the VMI state changes.
The violation reports are stored in the side blocks of
the block chain, which provides a full-proof means
of monitoring/auditing the VMIs. The VMIs can un-
dergo innumerable state changes, all of which are
well reflected in the blockchain that stores the entire
VMI metadata. Moreover, due to the inherent prop-
erty of immutability in blockchain,it can be assumed
that an adversary can not control a majority of the
computing power in the network, as a result avoiding
51 percent attack (Conti et al., 2018). The adminis-
trator/system admin level compromise also becomes
quite difficult in this environment due to the presence
of miners. Impersonification by the intruder also be-
comes a tough job here, as the private cryptographic
keys (Conti et al., 2018) represents each uniquepartic-
ipant of the network.
6 CONCLUSION
In this article, a formal model of blockchain based
VMI Monitor has been proposed that runs on an
Ethereum platform. Different components of the pro-
posed model have been described in detail. The Cloud
or the VMI operations are main concern in this work.
They have been elaborated and the different phases
of the entire procedure have been detailed. Further-
more, the smart contract for this particular problem
has also been designed along with the utility and role
of the various variables and functions. It has also been
established that the proposed system is better than the
previous scheme(Kazimet al., 2013) that has been im-
plemented to solve the same set of issues.
It has been observed that the smart contract based
system not only provides a proper VMI access con-
trol mechanism but also provides the scope of VMI
auditability and secured operations from the perspec-
tive of CSC as well (as observed that the side blocks
are checked when any CSC with high sensitivity is
dealt with). This methodology turns to be beneficial
not only for the CSP, who can now maintain a secured
Cloud infrastructure but also for the Cloud Service
Consumer (CSC) who can now deploy his applica-
tion/data in a VMI without the fear of getting his/her
data compromised.
The main overhead lies in the scalability of the
system with the number of VM transactions i.e. num-
ICISSP 2021 - 7th International Conference on Information Systems Security and Privacy
438
ber of CSC request. The more the number of re-
quest, more the size of the blockchain and with the
increased number of blocks the time required to final-
ize the same and add it to the main chain using the
PoS (Baliga, 2017) consensus method increases. An
effort has been made here (as mentioned in the previ-
ous sections) to balance the number of transactions by
following a user defined periodic update method. This
avoids unnecessary immediate transactions involving
blockchain write.
Future work is geared towards the development of
an automated tool based on the proposed methodol-
ogy. Another aspect that can be considered as fu-
ture work is enactment of the security issues related
to virtual machine instances for multi-provider feder-
ated clouds. The model, in its current form, considers
only single CSP set-ups.
REFERENCES
Abbadi, M. I. (2011). Clouds’ infrastructure taxonomy,
properties, and management services. In Advances in
Computing and Communications, Part IV, edited by
A. Abraham, J. L. Mauri, J. Buford, J. Suzuki, and S.
M. Thampi, Heidelberg: Springer-Verlag, 406-420.
Anonymous (2019). Local and state solidity vari-
ables and use of parameters. BitDegree Learn.
https://www.bitdegree.org/learn/solidity-variables,
Accessed 30 September 2019.
Baliga, A. (2017). Understanding blockchain consensus
models. InPersistent. https://www.persistent.com/wp-
content/uploads/2018/02/wp-understanding-
blockchain-consensus-models.pdf, Accessed April
2017.
Buyya, R., Yeo, C. S., Venugopal, S., Broberg, J., and
Brandic, I. (2009). Cloud computing and emerging
it platforms: Vision, hype, and reality for delivering
computing as the 5th utility. Future Generation com-
puter systems, 25(6), 599-616.
Conti, M., Kuma, E. S., Lal, C., and Ruj, S. (2018). A
survey on security and privacy issues of bitcoin. IEEE
Communications Surveys and Tutorials. 20(4), 3416-
3452.
Dannen, C. (2017). Introducing ethereum and solidity.
In: Dannen, C. (eds.) (Vol. 1), Berkeley: Apress, pp.
1–185. Springer Nature Switzerland.
Ferris, J. M. (2014). Red hat inc.: Load balancing in cloud-
based networks. U.S. Patent 8,849,971.
Halpin., H. and Piekarska, M. (2017). Introduction to se-
curity and privacy on the blockchain. In 2017 IEEE
European Symposium on Security and Privacy Work-
shops (EuroS and PW), pp. 1-3, IEEE.
Hashizume, K., Rosado, D. G., Fern´andez-Medina, E., and
Fernandez, E. B. (2013). An analysis of security is-
sues for cloud computing. In Journal of internet ser-
vices and applications. SPRINGER.
Hussein, R. K., Alenezi, A., Wills, G. B., and Walters, R. J.
(2016). A framework to secure the virtual machine
image in cloud computing. In 2016 IEEE international
conference on smart cloud (SmartCloud), pp. 35-40,
IEEE.
Kazim, M., Masood, R., and Shibli, M. A. (2013). Secur-
ing the virtual machine images in cloud computing.
In Proceedings of the 6th International Conference on
Security of Information and Networks, pp. 425-428.
Macrinici, D., Cartofeanu, C., and Gao, S. (2018). Smart
contract applications within blockchain technology: A
systematic mapping study. Telematics and Informat-
ics, 35(8), 2337-2354.
Marr, M. D. and Kowalski, M. P. (2014). Amazon tech-
nologies inc. scaling a virtual machine instance. U.S.
Patent 8,825,550.
Nakamoto, S. (2008). Bitcoin: A peer-to-peer electronic
cash system. https://bitcoin.org/bitcoin.pdf.
Sahli, H., Belala, F., and Bouanaka, C. (2014). Model-
checking cloud systems using bigmc. In 8th Inter-
national Workshop on Verification and Evaluation of
Computer and Communication Systems, pp. 25-33.
Schwarzkopf, R. (2015). Virtual machine lifecy-
cle management in grid and cloud computing.
https://doi.org/10.17192/z2015.0407.
Wan, Z. and Jiang, X. (2010). Hypersafe: a lightweight ap-
proach to provide lifetime hypervisor control-flow in-
tegrity. In IEEE Symposium on Security and Privacy,
pp. 380-395, IEE.
Wang, W., Hoang, D., Hu, P., Xiong, Z., Niyato, D., Wang,
P., Wen, Y., and Kim, D. I. (2019). A survey on
consensus mechanisms and mining strategy manage-
ment in blockchain networks. IEEE Access. 7, 22328-
22370.
Wei, J., Zhang, X., Ammons, G., Bala, V., and Ning, P.
(2009). Managing security of virtual machine images
in a cloud environment. In Proceedings of the 2009
ACM workshop on Cloud computing security, pp. 91-
96.
Wood, G. (2014). Ethereum: A secure decentralised gen-
eralised transaction ledger. Ethereum project yellow
paper, 151, 1-32.
Yann300 (2020). Remix documentation. Re-
lease 1, https://readthedocs.org/projects/remix-
ide/downloads/pdf/latest/.
Blockchain based Secured Virtual Machine Image Monitor
439