40th Annual IEEE Conference on Local Computer Networks LCN 2015, Clearwater Beach, Florida, USA

Software Defined Networking for Wireless Local

Networks in Smart Grid
Kemal Akkaya, A. Selcuk Uluagac and Abdullah Aydeger
Dept. of Electrical & Computer Engineering, Florida International University, Miami, FL, 33174 USA
[email protected], [email protected], [email protected]

Abstract—Emerging Software Defined Networking (SDN) tech- ways [12]. The data communication motivation necessitated
nology has provided excellent flexibility to large-scale networks in upgrading the existing Smart Grid network infrastructure with
terms of control, management, security, and maintenance. With different components such as home area networks (HANs),
SDN, network architectures can be deployed and maintained
with ease. New trends in computing (e.g., cloud computing, data neighborhood area networks (NANs) and wide-area networks
centers, and virtualization) can seamlessly be integrated with the (WANs). Each of these networks deploy thousands of network
SDN architecture. On the other hand, recent years witnessed devices that need to be managed continuously. Unfortunately,
a tremendous growth in the upgrade and modernization of this massive infrastructure requires additional labor and cost
the critical infrastructure networks, namely the Smart-Grid, for the utility companies who own these networks. Although
in terms of its underlying communication infrastructure. From
Supervisory Control and Data Acquisition (SCADA) systems to minimizing the management cost is one of the goals of the
Advanced Metering Infrastructure (AMI), an increasing number utilities, this cost will always be relevant as long as the
of networking devices are being deployed to connect all the local customers are served. In fact, the emerging SDN paradigm
network components of the Smart Grid together. Such large can provide excellent opportunity for reducing the network
local networks requires significant effort in terms of network management cost by integrating a software-based control that
management and security, which is costly in terms of labor
and hardware upgrades. SDN would be a perfect candidate can be flexible with respect to software upgrades, flow-control,
technology to alleviate the costs while providing fine-grained security patching, and quality of service (QoS). Nonetheless,
control of this critical network infrastructure. Hence, in this while a significant amount of work has been done in the SDN
paper, we explore the potential utilization of the SDN technology space, most of these efforts targeted the applications in the
over the Smart Grid communication architecture. Specifically, area of cloud computing, data centers and virtualization [5]
we introduce three novel SDN deployment scenarios in local
networks of Smart Grid. Moreover, we also investigate the and there is a need to adapt SDN for Smart Grid applications.
pertinent security aspects with each deployment scenario along This work is the result of such an effort to promote the use
with possible solutions. of SDN for various applications in the Smart Grid. Specifically,
Keywords: SDN; wireless local networks; smart grid; micro- we aim three different Smart Grid applications that rely on
grid; security a local wireless network infrastructure. 1)AMI applications
where meter data are collected via a mesh network that
I. I NTRODUCTION consists of smart meters and relays. Each of these equipment
The continuous growth of the Internet and the proliferation will have the ability to route the meter data through their
of smart devices and social networks pose new challenges routing tables. 2) SCADA Systems, which connects field de-
for networks on keeping up with the dynamicity of the vices such as relays, IEDs (Intelligent Electronic Devices),
hardware and software. In particular, the switches and routers PLCs (Power Line Communications), and PMUs (Phasor
that are involved in the transmission of the data from these Measurement Units) with the control center using redundant
networks and devices are typically developed in a vendor wireless connections. The control center is typically equipped
specific fashion, which makes hardware and software updates with routers and switches just like a data center. 3) Microgrid
a significant challenge. The emerging SDN technology is a Systems which integrates distributed power resources with the
solution to address such problems that can facilitate updates Smart Grid. The control and monitoring of these networks
to the hardware and software used on the networking devices require the deployment of network devices for collecting data
[5]. SDN enables splitting controls of networks and data flow about them.
operations. One of the major goals in SDN is to be able to This paper is organized as follows. In II, we provide some
interact with the switches and thus create an open networking background on SDN. In Section III, we describe how three
architecture for everyone. In this way, one can get a global Smart Grid applications can exploit SDN. Section IV explores
view of the entire network and will be able to make global potential security threats related to Smart Grid-enabled SDN.
changes without having to access to each device’s unique Finally, we conclude the paper in Section VI.
hardware.
II. BACKGROUND ON SDN
On the other hand, the existing power grid in the US is going
through a massive transformation to make it more reliable and SDN’s main motivation is to move the control of the lookup
connected with the ability to transfer data and power in two- tables inside the network devices to a separate location so

978-1-4673-6773-8/15/$31.00 ©2015 IEEE 826

that it can be controlled more easily. Specifically, this can be save time and energy. This idea of workload offloading can
described as separation of the packet forwarding and the way also be used for mobile devices which needs more powerful
how the forwarding tables are created and changed. These machines with certain security requirements [3]. Finally, SDN
two processes are assumed to be on separate layers, which are can be used for Internet research for testing certain ideas
referred to as data plane and control plane in SDN technology: without changing the current network [4].
Control Plane is also referred to as Network Operating System III. I NTEGRATION OF SDN WITH T HE S MART G RID
(NOS). This plane is supposed to supply network decisions
for packet forwarding or dropping. On the other hand, data In this section, we first explain the communication infras-
plane is responsible for data forwarding. SDN focuses on flow tructure for Smart Grid and then detail the applications that
forwarding idea instead of dealing with IP or MAC-address can benefit from the use of SDN.
based forwarding. A. Smart Grid Networking Architectures
Main problem in the traditional networks was updating Before moving into the description of how SDN can be
the network elements, which requires too much effort and used for Smart Grid applications, we first briefly explain the
time [5]. By creating a programming interface to be able to existing Smart Grid network infrastructure. Basically, there
update network elements from a center (i.e., at the control are three major components in a Smart Grid infrastructure
plane) via SDN, such complexity in network management can [12]: 1) HANs that mainly connect home devices with the
be eliminated. An illustration of how SDN reorganizes the smart meters; 2) NANs which collect smart meter data from
network architecture with respect to the current ones are is houses; and 3) WANs which provide long-haul communication
illustrated in Fig. 1. As can be seen from this figure, the control with the utility control centers using various technologies
plane is typically controlled by a single center through some including cellular ones. A typical Smart Grid networking
applications. infrastructure showing possible infrastructure for generation,
transmission, and distribution components of power grid is
Switch A Switch C
depicted in Fig. 2. Under this infrastructure, many applications
Control A Control C could run simultaneously. In this work, we focus on three
Data A Data C
different applications that utilize this networking infrastructure
Switch B Switch D
at different levels as discussed below.
Control B Control D
Data B Data D
Control/Operations Centers

Communications
Network Utility
Neighborhood
Wide Area Area Network
Network
Computer 1 Computer 2
Home Area
Network

(a)
Electric
Vehicle
Apps Updates
Management Apps Security
Power Field Field Field Smart Home
Plant Devices Devices Devices Meter Appliances
Control
Plane Network Operating System

OpenFlow Generation Transmission Distribution Customer Premises

Switch CPU
Protocol
Match Action Stats

Flow Table
Switch A Switch C
Fig. 2: Smart Grid multi-tier communication network [12]
Data A Data C

Switch B Switch D
B. Motivation for SDN-Enabled Smart Grid
Data
Plane Data B Data D
Utilities and energy companies own and operate network
components that are part of the Smart Grid communication
infrastructure. In some cases, they also lease services from
Computer 1 Computer 2 telecommunication companies or third party cloud services.
In any case, the management of these networks is a great
(b) challenge due to the scale. Furthermore, due to devices and
Fig. 1: Current (a) vs SDN-based (b) Networks applications from different vendors, the equipments may not
be interoperable. Therefore, the utilities will need to deal
Today, SDN is used for a variety of networking settings. For with equipment maintenance and software upgrade, which
instance, it is used in cloud computing and data centers for is costly and laborous. SDN is a promising solution for
ease of network management and control of virtual machines the aforementioned problems due to following advantages it
(VMs) [18]. It can connect multiple data center networks by presents:
eliminating the problems of proprietary architectures. In this • SDN adopts open standards and introduces technology
way, the workload can move from one network to another to abstraction, which provides a vendor-agnostic approach

827
 to configuring and maintaining various types of network
elements that are common in Smart Grid.
SDN Control
• Hardware virtualization through SDN eases the burden Plane

of managing different networks while using resources WAN

efficiently.
• Due to its holistic view of network, the SDN-based
Gateway
network provides superior control of delay and jitter in Utility Control Center

the network which is crucial for SCADA systems in terms

of power and load state estimation and control.
• SDN’s bandwidth-on-demand capabilities can also create OpenFlow

802.11s

opportunities to increase revenue through accelerated 802.15.4g

service velocity in cases where the utility also serves

Fig. 3: Proposed SDN for Smart Grid NANs
as a communication service provider in the coverage
area. More and more utility companies are functioning
as service providers in rural areas. consuming information about the status of some aspect of
IV. P ROPOSED SDN D EPLOYMENT S CENARIOS IN S MART the substation. A standard called IEC 61850 is also used for
G RID substation automation [2]. Currently, these systems are not
only scalable, but also their sensing coverage is very limited.
A. SDN-enabled NANs
Proper configuration and maintenance of IED communication
Smart Grid’s NAN is mainly used for AMI applications. requires significant effort. The network complexity further
While there has been some wired options for building these increases with the uses of other protocols.
communications, recent implementations solely targeted wire-
However, with the modernization of the power grid, there
less solutions that depended on different standards such as
will be opportunities to upgrade these systems based on a mix
IEEE 802.15.4g, IEEE 802.11s, RF-Mesh, and other propri-
of wireless and wired infrastructure through the deployment
etary mesh networks [12]. This in turn creates a wireless local
of a large number of modern PMUs. For instance, the design
network that can be used in a particular neighborhood. As long
and use of wireless PMUs have already begun [10]. One of
as different vendors’ products support OpenFlow, a NAN using
the recent works proposed using these wireless PMUs within
a mixture of these standards can be easily controlled and re-
an SDN architecture so that the network administrator would
tasked through a standard network control script programming.
have a global view of the power grid computer network,
In most cases, Smart Grid operators prefer exchanging the
which makes it is easier to manage PMU telemetry traffic
information among different NANs in order to get a better
compared to a traditional IP network [7]. Again, this creates
load state estimation. Therefore, being able to control such a
a perfect example of a local network within a substation that
network in a centralized manner for load balancing, security
is supplemented by wireless communication.
and QoS services is very valuable. While SDN can provide
this novelty, there are still challenges that would require some There are also opportunities to reorganize the elements
research to enable the use of SDN in wireless environments. of these SCADA systems especially in terms of exploiting
One of these challenges is the performance of the central- efficient ways to eliminate the complexities of multicasting and
ized control. As opposed to wired networking interface among broadcasting. Massive amount of data is transmitted through
the controller and SDN switches, this will not be the case in these broadcasts or multicasts. Current architecture is a cen-
wireless-mesh based NANs due to the scalability of the AMI. tralized one with hub-and-spoke model which is inadequate to
Therefore, the control will be through wireless communication address too many broadcasts or multicasts. Researchers strived
and most probably using multi-hopping (see Fig. 3). If the to address this issue by using middleware approaches in the
same channel is used for data communications, then this may past [6]. The idea in these approaches are to implement publish
create a lot of interference. While some of the very recent and subscribe mechanisms where the data sources publish data
works investigated this issue for wireless mesh networks using and the brokers in the middleware (at the application layer)
[11], these works do not directly apply to AMI NANs where are responsible for delivering this data to subscribers within
the scale is larger and the variety of nodes is significant in their QoS requirements. This is in a way a sort of group
terms of used hardware/software. This suggests investigating communications among publishers and subscribers. However,
the feasibility of distributed control in SDN-based NANs. since this was implemented at the application layer, it is
not only slower, but also not flexible in terms of hardware
B. SDN-enabled SCADA requirements. SDN can be used to redesign this middleware
SCADA systems were designed to collect data from field by including the control plane in the middleware, but at the
devices such as PLCs, PMUs, IEDs at substations in real- network layer [20]. Another advantage of such an SDN-based
time and do control decision at the control center in terms control is the ability to perform traffic engineering, which
of reliability and quality of the power [2]. A substation cannot be done with layer-2 switches using spanning tree-
contains hundreds of different IEDs, each generating and/or based routing [1].

828
C. SDN-enabled Microgrid
A microgrid is a miniaturized version of power grid which
can supply electrical load of small communities such as
university campuses, malls, camps etc. It includes numerous
Distributed Energy Resources (DERs) (e.g., photovoltaic sys-
tems, micro-combined heat and power systems (µCHP), and
electric vehicles (EVs)), load, storage, and protection devices
that are controlled by a central controller. Microgrids are
becoming viable options for saving energy and generating Fig. 4: A threat model for the SDN-enabled smart grid
clean power along with their reliable electrical services. For
instance, Univ. of California San Diego setup a micro grid for any device such as IEDs, PMUs, PLCs, Smart Meters could
its own campus which saved them $850K a month [17]. be valuable targets for potential malicious activities.
The major issue with these microgrids is the risk of rapid In software-specific threats, the attackers aim to exploit the
changes that may cause instability and eventually collapse in vulnerabilities associated with the networking protocols, soft-
the system. Therefore, it is crucial to perform fine-grained real- ware suits (IEC 61850, IEEE C37.118 Syncrophasor Protocol,
time monitoring and control. This can only be achieved via Openflow of SDN) that run in the smart grid.
reliable communications that can provide QoS in support of Finally, depending on the identity of the attacker, i.e.,
low data latency, packet prioritization and traffic engineering. whether an attacker is a legitimate member of the network
These features can be supported via the SDN technology to during an attack or not, she can be defined as insider or out-
stabilize and optimize system operation. Another opportunity sider attacker. Insiders are more dangerous than the outsiders
for the use of SDN in microgrids is on the problem of DER as they have more knowledge about the internal architecture
management and aggregation. Basically, the DERs are grouped of the SDN-enabled smart grid.
together for different purposes and any mobility related group In reality, there is no hard line between these attacking
changes will affect the system. Most of the DERs are expected models and they complement each other because an insider
to connect via a wireless link (e.g., EVs) which will create a could be a passive attacker trying to exploit IEC 61850 on an
local network consisting of wireless DERs and other wired IED in the SDN-enabled smart grid. The threat model for the
IEDs connected with the microgrid. Currently, DER mobility SDN-enabled smart grid is presented in Fig. 4.
and group management are done at the application layer. The B. Desired Security Mechanisms
complexity can be reduced by exploiting SDN capabilities that Desired security mechanisms are usually defined by the na-
will be implemented at the network layer [20]. tional and international standardization bodies (e.g., National
V. S ECURITY OF SDN-E NABLED S MART G RID Institute of Standards and Technology, International Telecom-
munication Union (ITU)) and are used by many researchers
In this section, we discuss the security of the SDN-enabled and practitioners who aim to develop secure systems. In this
smart grid. First, we articulate the threat model, then we list sub-section, we use the security architecture suggested by the
the desired security services for the SDN-enabled smart grid. ITU’s Recommendation X.800 [19] documentation, which is
A. Threat Model referred to as the Security Architecture for Open Systems
Interconnect (OSI) as our guideline in addressing the threats
Conceptually, the threats to the SDN-enabled smart grid discussed in the previous sub-section.
could be listed from four different complementary perspec- Confidentiality: Confidentiality refers to the protection of
tives: (1) Method-specific; (2) target-specific; (3) software- the exchanged content (e.g., gathered data, reports, commands)
specific; and (4) identity-specific. among the components of the smart grid such as IEDs,
Method-specific threats define how the threats are executed. PMUs, PLCs, Smart Meters. A malicious entity which has
The method-specific threats can be either passive or active. In the privilege to access the content, should not be able to
the passive method, the attacker only monitors (or eavesdrops), decode the exchanged messages in the network. Confidentially
records the communication data occurring in the SDN-enabled also entails the protection against any unintended information
smart grid, and analyzes the collected data to gain meaningful leakage from the applications, controllers, and devices within
information. In the active one, the attacker tries to send fake the SDN-enabled smart grid. This is particularly important
authentication messages, malformed packets, or replay a past because the data generated and collected by the smart grid
communication to the components of the the SDN-enabled equipment, e.g., PMUs, IEDs are very periodic in its nature.
smart grid. As passive threats are surreptitious, it is harder Data forwarding policies or flow rules associated with the
to catch their existence. However, it is easier to catch the collected data may be discovered with simple timing or
existence of an active attacker, but its damage to the smart side-channel analysis. Similarly, an increased delay for the
grid can be relatively higher than the passive threats. establishment of a new flow rule in response to an incoming
Target-specific threats classify the attacks according to packet can inform a potential attacker about the behavior of the
which device the threats target. In an SDN-enabled smart grid, OpenFlow controller within the SDN-enabled smart grid. This

829
unintended information disclosure from data plane devices, stale because the nature of the messages in the smart grid
applications, flows, controllers should also be considered as are very time-sensitive. Within the SDN-enabled smart grid,
part of any confidentiality service. modification of the flow rules or insertion of new Openflow
Traditionally, confidentiality can be provided by adopt- rules [8] by adversaries can cause severe damage to the healthy
ing either symmetric and asymmetric key-based encryption operations of the smart grid.
schemes [16]. In symmetric encryption, one key is utilized Integrity is usually provided by appending the cryptographic
among the PMUs, PLCs, smart meters, IEDs, applications, digest of the message content to the message itself [16]. When
flows, network controllers. Examples of symmetric encryption the PMUs, PLCs, smart meters, IEDs, applications, network
that can be utilized for the smart grid include AES, RC4. On controllers receive the message, they can check to see if the
the other hand, in asymmetric encryption, a pair of two keys digest of the content matches the digest they computes on
(aka public and private) are utilized among the communicating their end. If the digests match each other, then the message is
components of the smart grid. RSA and ECC are the two most deemed legitimate and not to have changed from its original
important examples of asymmetric encryption that could be content. Content digests in integrity are usually created with
deployed. Moreover, the maturing state-of-the art encryption the usage of hashing algorithms. There are several hashing
mechanisms based on fully-homomorphic-encryption could be algorithms such (e.g., MD5, SHA-2) in use today, which do
utilized for specifically preserving the privacy of the flows. not require the presence of keys unless they are specifically
Authentication: Authentication involves guaranteeing the designed to work with keys like keyed-hashing (e.g., HMAC,
genuineness of the communication among the devices in the CMAC). Alternatively, integrity can be provided as part of
data plane, controllers, and the applications. An authentication a digital authentication mechanism utilizing symmetric and
mechanism verifies if the exchanged information stems from asymmetric encryption techniques. For instance, the last block
the legitimate participants of the SDN-enabled smart grid of the encrypted data in AES can be appended to the message
because a malicious entity (e.g., a compromised IED) may that would be sent as the integrity code. In a similar fashion,
be able to inject counterfeit content or resend the same a private key in the asymmetric encryption techniques (e.g.,
content into the SDN-enabled smart grid. More specifically, an RSA, ECC) can be used to provide the integrity code appended
adversarial smart grid application may attempt to insert new to the message.
flow rules that may circumvent flow rules imposed by other Access Control: With access control, unauthorized use of a
applications [14]. Adversaries may also insert new rules to resource in the SDN-enabled smart grid is prevented. Access
damage the system by influencing the state estimation, which control addresses which participant of the smart grid reaches
is crucial to evaluate the demand. which content or service. For instance, IEDs should not be
Authentication can fundamentally be provided based on allowed to have the privileges of PMUs. Proper security
three factors [16]: (1) Knowledge factor: the proof of the measures must prevent any unauthorized SDN controller ac-
knowledge of some secret (e.g., passwords) is provided to the cess. An unauthenticated application might try to access to
authenticator. Symmetric, asymmetric key-based encryption resources for which it does not have exclusive privileges. Or,
schemes and hashing algorithms can all be utilized as part an authenticated application, IEDs, PMUs, PLCs, and Smart
of the authentication mechanism with the knowledge factor. Meters may abuse its privileges.
(2) Possession factor: authenticator verifies the claimant using Access control is usually achieved through four different
the credentials provided by a specialized hardware. Electronic methods [16]: (1) discretionary access control (DAC); (2)
cards, smart cards, smart tokens physically owned by the mandatory access control (MAC); (3) role-based access con-
claimant can be utilized and integrated with the SDN-enabled trol (RBAC); and (4) attribute-based access control (ABAC).
smart grid devices and applications. (3) Identity factor: the In DAC, access control decisions are made based on the
authenticator utilizes features uniquely identifying in the ver- exclusive rights that are set for the flows, applications, IEDs,
ification of the claimant. Both static or dynamic patterns that PMUs, PLCs, and Smart Meters. An entity in DAC can enable
can identify the devices and applications can be utilized. For another entity for accessing resources. In MAC, access control
instance, behavioral information from the SDN-enabled smart function considers the criticality of the resources and the rights
grid devices and applications such as communication patterns, of the flows, applications, IEDs, PMUs, PLCs, and Smart
timing patterns, delays can all be utilized [9] as part of this Meters on the resources. In MAC, an entity can not enable
authentication method. Within the SDN-enabled smart grid, another entity for accessing the resources. In RBAC, access
all of these authentication techniques can be individually or control decisions are based on the roles created within the
a combination of one or more of the techniques could be the SDN-enabled smart grid. A role can include more than
adopted. If more than one factor is utilized, the authentication one entity e.g., the flows, IEDs. Moreover, a role defines the
is called multi-factor authentication. capabilities what the entities can do or not do within a certain
Integrity: Integrity refers to the capability to detect detect role. Finally, in ABAC, the access control decisions are based
if the exchanged content between the communicating devices on the features of the flows, applications, IEDs, PMUs, PLCs,
of the smart grid have been altered or not. Furthermore, and Smart Meters, resources to be accessed, and environmental
the integrity service involves ensuring that the exchanged conditions.
content is not deleted, replication of old data, counterfeit, or Availability: Due to the threats to SDN-enabled smart grid,

830
some portion of the grid or some of the functionalities or local networks. For instance, the experience in an SDN-based
services provided by the grid could be damaged and unavail- NANs can be adapted in a community wireless mesh network.
able to the participants of the grid. For instance, some PLCs In the same manner, vehicular ad hoc networks can also benefit
could be compromised and they could cease functioning. A from DER deployments in microgrids.
Denial-of-Service (DoS) type attack [13] can overflow the R EFERENCES
communication link of the SDN controller-switch [14]. SDN
flow switch tables can be flooded by fake entries. In a similar [1] Ian F. Akyildiz, Ahyoung Lee, Pu Wang, Min Luo, and Wu Chou. A
roadmap for traffic engineering in sdn-openflow networks. Comput.
fashion, a centralized SDN controller can be a single point Netw., 71:1–30, October 2014.
of failure. Moreover, recent technological advances enabled [2] Kenneth C Budka, Jayant G Deshpande, Tewfik L Doumi, Mark Mad-
the integration of the wireless technologies (e.g., ) into the den, and Tim Mew. Communication network architecture and design
principles for smart grids. Bell Labs Technical Journal, 15(2):205–227,
smart grid infrastructure. In such cases, adversaries may jam 2010.
the wireless medium, effectively hampering all the communi- [3] R. Hill N. Bessis C. Baker, A. Anjum and S. L. Kiani. Improving cloud
cations. Thus, availability service ensures that the necessary datacenter scalability, agility and performance using openflow. 2012.
[4] P. F. Rosa F. de O. Silva, J. H. de S. Pereira and S. T. Kofuji.
functionalities or the services provided by the SDN-enabled Enabling future internet architecture research and experimentation by
smart grid are always carried out, even in the case of attacks. using software defined networking. 2012.
Usually, the smart grid includes redundant components in [5] Qi Hao Fei Hu and Ke Bao. A survey on software-defined network and
openflow: From concept to implementation. 2014.
their infrastructure. This is to ensure the continuous operation [6] Harald Gjermundrod, David E Bakken, Carl H Hauser, and Anjan Bose.
during failures. In a similar fashion, the SDN-enabled smart Gridstat: A flexible qos-managed data dissemination framework for the
grid can be designed with such redundancy to achieve the power grid. Power Delivery, IEEE Transactions on, 24(1):136–143,
2009.
availability service. [7] A. Goodney, S. Kumar, A. Ravi, and Y.H. Cho. Efficient pmu networking
Accountability: With accountability (aka non- with software defined networks. In Smart Grid Communications (Smart-
repudiation [15]), the SDN-enabled smart grid ensures GridComm), 2013 IEEE International Conference on, pages 378–383,
Oct 2013.
that a device or a software component (e.g., applications, [8] R. Kloti, V. Kotronis, and P. Smith. Openflow: A security analysis. In
IEDs, PMUs, PLCs, and Smart Meters) can not refute the Network Protocols (ICNP), 2013 21st IEEE International Conference
reception of a message from the other device or application on, pages 1–6, Oct 2013.
[9] Wenyi Liu, A.S. Uluagac, and R. Beyah. Maca: A privacy-preserving
or the sent of a message to the other device or application in multi-factor cloud authentication system utilizing big data. In Com-
the communication. puter Communications Workshops (INFOCOM WKSHPS), 2014 IEEE
Conference on, pages 518–523, April 2014.
Accountability can be provided as a service bundled inside [10] Brian Miller. Concept for next generation phasor measurement: A low-
authentication and integrity. For instance, a digital signature cost, self-contained, and wireless design. Master’s thesis, University of
scheme (DSS) [15], which is based on utilizing encryption Tennessee, 2010.
[11] V. Nascimento, M. Moraes, R. Gomes, B. Pinheiro, A. Abelem, V.C.M.
methods would address accountability. Additionally, proper Borges, K.V. Cardoso, and E. Cerqueira. Filling the gap between
auditing mechanisms and logs should be utilized to provide software defined networking and wireless mesh networks. In Network
accountability in the SDN-enabled smart grid. and Service Management (CNSM), 2014 10th International Conference
on, pages 451–454, Nov 2014.
VI. C ONCLUSION [12] Nico Saputro, Kemal Akkaya, and Suleyman Uludag. A survey of
routing protocols for smart grid communications. Computer Networks,
In this work, we introduced how the emerging SDN 56(11):2742 – 2771, 2012.
paradigm could be considered as a viable technology for the [13] L. Schehlmann, S. Abt, and H. Baier. Blessing or curse? revisiting
security aspects of software-defined networking. In Network and Service
Smart Grid communication architecture, which is currently Management (CNSM), 2014 10th International Conference on, pages
under massive modernization effort by the utility providers. 382–387, Nov 2014.
We discussed how flexibility and ease of control, management, [14] S. Scott-Hayward, G. O’Callaghan, and S. Sezer. Sdn security: A survey.
In Future Networks and Services, 2013 IEEE SDN for, pages 1–7, Nov
security, and maintenance provided by the SDN technology 2013.
could make a compelling case for applying SDN in three [15] William Stallings. Cryptography and Network Security: Principles and
unique smart grid deployments that will mimic a wireless Practices (3rd edition). Prentice Hall, 2003.
[16] William Stallings and Lawrie Brown. Computer Security: Principles
local network: Specifically, we focused on SCADA systems, and Practice (3rd edition). Prentice Hall, 2015.
AMI, and Microgrid Systems and discussed how an increasing [17] Ucsd microgrid.
number of smart grid devices that are being deployed to [18] K. Kannan V. Mann, A. Vishnoi and S. Kalyanaraman. Crossroads:
Seamless vm mobility across data centers through software defined
connect all the components of the Smart Grid together could networking. 2012.
benefit from the SDN. Furthermore, we articulated potential [19] ITU-T Recommendation X.800. Security architecture for open systems
security threats that could arise in an SDN-enabled smart grid interconnection for ccitt applications. 1991.
[20] Jianchao Zhang, Boon-Chong Seet, Tek-Tjing Lie, and Chuan Heng
and provided some potential solutions to alleviate the threats. Foh. Opportunities for software-defined networking in smart grid. In
Applying the maturing SDN technology into the smart grid Information, Communications and Signal Processing (ICICS) 2013 9th
infrastructure presents ample unique research challenges in International Conference on, pages 1–5, Dec 2013.
security and networking to engineers and scientists and to
the best of our knowledge this is the first work exploring
those challenges. We would like to emphasize that the ideas
mentioned here will be perfectly applicable to other wireless

831