An Architecture for

Software Defined Wireless Networking

Carlos J. Bernardos, Antonio de la Oliva, Hao Jin

Pablo Serrano, Albert Banchs InterDigital Communications, Inc
Universidad Carlos III de Madrid 781 Third Avenue, King of Prussia
Avda. Universidad, 30 Pennsylvania, USA
Leganes, Spain [email protected]
{cjbc, aoliva, pablo, banchs}@it.uc3m.es

Luis M. Contreras Juan Carlos Zúñiga

Telefónica I+D InterDigital Communications, LLC
c/ Don Ramón de la Cruz, 82-84 1000 Sherbrooke Street West
Madrid, Spain Montreal, Canada
[email protected] [email protected]

Abstract— Software Defined Networking (SDN), characterized machinery [2]. As a natural next step, mobile network
by a clear separation of the control and data planes, is being operators need to re-think how to evolve their existing
adopted as a novel paradigm for wired networking. With SDN, network infrastructures and how to deploy new ones to
network operators can run their infrastructure more efficiently, address the challenges posed by the increasing customers'
supporting a faster deployment of new services while enabling demands, as well as by the huge competition among operators.
key features such as virtualization. In this article, we adopt an All these changes are triggering the need for a modification in
SDN-like approach applied to wireless mobile networks that will the way operators and infrastructure providers operate their
not only benefit from the same features as in the wired case, but networks, as they need to significantly reduce the costs
also will leverage on the distinct features of mobile deployments incurred in deploying a new service and operating it.
to push improvements even further. We illustrate with a number
Some of the mechanisms that are being considered and already
of representative use cases the benefits from the adoption of the
adopted by operators include: sharing of network
proposed architecture, which is detailed in terms of modules,
interfaces and high-level signaling. We also review the ongoing
infrastructure to reduce costs, virtualization of core servers
standardization efforts, and discuss the potential advantages, running in data centers as a way of supporting their load-
weaknesses and the need for a coordinated approach. aware elastic dimensioning, and dynamic energy policies to
Keywords — SDN, Wireless, Standardization, ONF, reduce the monthly electricity bill. However, this has proved
OpenFlow, OmniRAN, Access, Backhaul to be tough to put in practice, and not enough. Indeed, it is not
easy to deploy new mechanisms in a running operational
network due to the high dependency on proprietary (and
I. INTRODUCTION sometime obscure) protocols and interfaces, which are
The telecommunications sector is experiencing a major complex to manage and often require configuring multiple
revolution that will shape the way networks and services are devices in a decentralized way.
designed and deployed for the next decade. We are witnessing Building on the revolutionary forward thinking in computer
an explosion in the number of applications and services networking, software defined networking (SDN) is currently
demanded by users, which are now really capable of accessing being considered as an alternative to classic distributed
them on the move. In order to cope with such a demand, some approaches based on highly specialized hardware executing
network operators are now following a cloud computing standardized protocols. Up to now, most of the key use cases
paradigm, enabling the reduction of the overall costs by used to present the benefits of the SDN paradigm have been
outsourcing communication services from specific hardware limited to wired environments (e.g., Google uses SDN in its
in the operators’ core to server farms scattered in datacenters. datacenters [3]).
These services have different characteristics if compared with In this article we analyze the potential of applying the SDN
conventional IT services that have to be taken into account in paradigm to mobile wireless networks. First, we identify use
this cloudification process [1]. cases where a wireless SDN approach could bring additional
Virtualization of functions also provides operators with tools benefits. Then, we derive the main characteristics of a wireless
to deploy new services much faster, as compared to the SDN mobile operator`s architecture, paying special attention
traditional use of monolithic and tightly integrated dedicated to the main functions and interfaces. In order to illustrate the
operation of the proposed mobile wireless SDN framework, A key requirement to deploy an SDN architecture, such as the
we introduce the high-level interactions required between the one defined above, is to standardize the interface to control the
defined functions to support an example use case of interest mobile devices. This can be done with OpenFlow [5], which is
for operators. We finish this article with a review of current a standardized interface between the control and forwarding
standardization efforts and trends in this arena, and we then layers of the SDN architecture. The vendor-agnostic nature of
elaborate on the need for specific actions towards the OpenFlow facilitates the integration of heterogeneous devices
standardization of what we call Software Defined Wireless in a common way, simplifying the operation of multi-vendor
Networking (SDWN). infrastructures, which are typically found in commercial
telecom networks. It allows accessing and modifying the
II. BACKGROUND: SDN, OPENFLOW, CAPWAP AND RE- forwarding plane of network devices such as switches and
CONFIGURABLE WIRELESS DEVICES routers by the definition of specific rules for matching packet
Software Defined Networking is a networking paradigm [4] flows against a selection of layer-2 to layer-4 packet header’s
that separates the control and the data forwarding planes. Such field values and the flows’ ingress port number. These rules
separation allows for quicker provisioning and configuration take the form of entries in forwarding flow tables residing in
of network connections. With SDN, network administrators the network devices.
can program the behavior of both the traffic and the network It should be noted that this separation of control and data plane
in a centralized way, without requiring independently for the switching fabric also exists, to some extent, in the
accessing and configuring each of the network’s hardware wireless domain. Indeed, the IETF standardized several years
devices. This approach decouples the system that makes ago the Control And Provisioning of Wireless Access Points
decisions about where traffic is sent (i.e., control plane) from (CAPWAP) protocol [6], which centralizes the control in
the underlying system that forwards traffic to the selected wireless networks. In principle, CAPWAP is technology
destination (i.e., data plane). Among other advantages, this agnostic and requires specific bindings for each considered
simplifies networking as well as the deployment of new access standard, although so far only the binding for 802.11
protocols and applications. In addition, by enabling has been defined. Radio configuration is expressed in terms of
programmability on the traffic and the devices, an SDN management information base elements included in the
network might be much more flexible and efficient than a standard, such as the operating channel or the transmission
traditional one. power, but also the beacon interval or the contention
Figure 1 shows a logical view of the commonly accepted SDN parameters used by the medium access scheme. With
reference architecture [4]. In this architecture, the intelligence CAPWAP, control frames are delivered to a central controller,
is centralized in software-based SDN controllers, which have a which is responsible for MAC layer control, in a way that can
global view of the network and are capable of controlling, in a be easily related to the way OpenFlow delivers to the
vendor-independent way, the network devices. These network controller information about new incoming flows.
devices are no longer required to implement and understand Along the same lines, but in a more visionary approach, a
many different network protocols standards; instead, they can novel paradigm for the re-programming of wireless interfaces
provide such functionality by accepting instructions from SDN has been proposed in [7]. In this vision, wireless nodes
controllers. This saves a lot of time and resources, as the execute a Wireless MAC Processor, in charge of running
network behavior can be easily controlled by programming it "MAClets", i.e., programs specifying the MAC protocol. In
in the centralized controllers, rather than using custom this way, the central controller can dynamically upload the
configurations in many different devices scattered across the protocol to use at a given point in time, e.g., changing from
network. CSMA/CA to TDMA-based access when the traffic load
increases.
There have also been some efforts looking at the use of SDN
in mobile networks, such as MobileFlow [8], which proposes
an SDN approach for the core network. However, the
proposed solution does not provide an integrated vision
including the wireless access.
While efforts in SDN so far have mostly focused on wired and
core networks, we believe that the adoption of a similar
concept for wireless access and backhaul environments can be
even more beneficial. Indeed, the control plane of wireless
networks is more complex than the one of wired networks, and
therefore higher gains can be achieved from the increased
flexibility provided by an SDN approach.
III. USE CASES
Before addressing the design of an SDWN architecture, in this
section we describe some use cases in which the adoption of a
Figure 1. SDN reference architecture
software defined wireless networking approach brings instances sharing the network, e.g., a change in the
significant advantages. Rather than identifying an exhaustive configuration of the virtual infrastructure must not affect the
set of use cases, the purpose of these use cases is simply to rest of the instances running on top of the real deployment.
illustrate some of the potential advantages of adopting an SDN Another key issue is the allocation and sharing of network
approach in a mobile wireless environment. resources, considering both time resolution and isolation. This
not only prevents resource wastage due to the coordination
A. Virtualization
and sharing, but also allows gaining from statistical
Current deployments support virtualization to some extent, but multiplexing, so the use of this solution is substantially more
existing network devices and mechanisms are not designed to efficient than having independent deployments.
support dynamic re-configuration required for timely and Note that SDN is independent from and complementary to
efficient sharing of resources. More specifically, although another notable virtualization initiative, the Network
servers can be virtualized and allocated to different physical Functions Virtualisation (NFV) [2]. While SDN focuses on the
resources in almost real time, the paths communicating these virtualization of network devices, NFV aims to enable the
virtual appliances with the rest of the world still require virtualization of network services and functions, such as NAT,
manual installation, or interaction with protocols which are not firewall and cellular core functions, so that the time to deploy
designed for dynamic and fast responses. Hence, current services can be shortened and operator CAPEX/OPEX can be
adaptation mechanisms in network technologies are seen as reduced. One example of the mutual-benefit between NFV and
one of the bottlenecks for the general deployment of SDN is that, on the one hand, NFV may improve the
virtualized infrastructures. Previous approaches for core efficiency and flexibility of SDN’s control plane services; and
network virtualization (e.g., PlanetLab1 or GENI2) have on the other hand, SDN may ensure the delivery and quality of
implemented different overlay networks so researchers can run the network traffic between NFV’s virtualized functions.
their experiments by time-sharing access to resources.
However, these approaches operate in a coarse time scale, B. QoE-aware network operation
need manual planning and dimensioning, and lack the required Current networks are provisioned and operated towards
timeliness for their operation in production networks. These providing a certain level of Quality of Service (QoS).
approaches differ from actual trends in virtualization for the However, this does not always ensure a minimum “Quality of
access network, more focused on sharing and enforcing the Experience” (QoE) to the user. The QoE of a service is
radio and transport network resources among different determined by roughly three factors: the service architecture
operators. (e.g., server capabilities, caches, and their location), the core
Following this trend, the adoption of SDN (with relatively network performance, and the service provided at the
mature technologies such as, e.g., OpenFlow or ForCES3) “wireless last mile”, i.e., the combination of the wireless
should improve support for timely and efficient virtualization link(s) – including the backhaul – and the capabilities of the
of a wireless network, but there are some challenges that need terminal. With current architectures, a service provider has to
to be successfully tackled. First, in order to provide the anticipate its needs in terms of infrastructure, negotiate an
required flexibility in terms of network topology and SLA based on these estimations, and at most try to adapt to
architecture demanded by virtualization applications, an SDN users’ experiences, in a coarse time-scale, e.g., by changing
network must be able to implement a wide set of control logics the encoding of the video being served (as YouTube does). It
that are simultaneously applied to the same set of physical is clear that such a scenario precludes an efficient use of
resources. The support of several different control logics on resources, as the service provider has no mechanism to react in
the same network raises scalability and compatibility issues. a timely manner to the changing conditions, because the
For example, each of the control logics may be working on top service provider has limited indicators of the user’s
of a different realization of the network, each of them performance in real time, nor the ability to quickly deploy
requiring fast reaction upon changes in the underlying more architectural elements or improve the SLA with the
physical infrastructure. This challenge calls for a scalable network providers.
network orchestration mechanism, which coordinates SDN Furthermore, mobile networks intrinsically present a need for
control and data plane operations and resolves any contentions integrating QoS objectives in the radio part (i.e., service layer)
between different control logics. and the backhaul network (i.e., transport layer) [9]. This drives
In addition, the different control logics must be able to work in the necessity of dynamically orchestrating resources in both
an isolated way. This must be implemented in two different layers for providing a uniform and efficient QoE.
planes. On the one hand, traffic from a certain instantiation of The use of a software defined wireless networking architecture
a virtual operator must be isolated from traffic belonging to a would allow the network to offer to the service provider an
different virtual operator, for security and privacy reasons. On API to control how the networks behave to serve the traffic
the other hand, changes performed in the virtual infrastructure that matches a certain set of rules (of course, the degree of
of an operator must be isolated from the rest of virtual control would depend on the agreements between network
operators and service providers and the kind of requested
1
control). Furthermore, through this API the provider will be
https://www.planet-lab.org/
2
http://www.geni.net/
also able to dynamically change the forwarding paths of the
3
http://datatracker.ietf.org/wg/forces/charter/ flows (in both directions), so traffic transverses
opportunistically deployed middleware, which can for instance important; second, by being aware of the service experienced
serve as data caches, or video transcoders. Finally, the by the user, thus timely adapting following his/her
provider, now acting as a true service composer, can use the preferences; third, by supporting traffic on- and off-loading,
API to change the behavior of the “wireless last mile” in three based on these preferences and the availability of different
ways: first, by dynamically prioritizing traffic at the last hop, communication links (each with a different performance vs.
so in case of poor wireless conditions, some packets (e.g., I cost trade-off). In this way, the provider personalizes the
frames of a video stream) are provided with a better service operation of the network after user’s behavior and preferences.
than others (e.g., B frames), because are marked as more

Figure 2. SDN-based mobile network architecture

The above requires the design of customer-oriented traffic how to select and opportunistically use the heterogeneous
management services, which are able to connect different available accesses is not a trivial one.
applications to different technologies and to adapt content to Additionally, the network might want to keep control on how
network conditions and available resources. By considering the traffic (on an application granularity level) is delivered to
the specifics of each application, a differentiated service can the mobile terminal. This involves, for example, selective and
be provided to different flows. For example, quality and/or opportunistic traffic offloading. By using an SDWN solution,
timely delivery of a given flow can be balanced against for an API could be offered to external parties (such as service
extra capacity for a different flow with more stringent delay providers), so they can influence the decision of which access
requirements. This requires the ability to design and technology is used to deliver a certain type of traffic to a
coordinate offloading/seamless mobility mechanisms across specific mobile terminal or group of users. This particular
the heterogeneous (technology and operator) networks. scenario could also benefit from enabling the programmability
of the mobile node as well, for example to enable easy control
C. Network access selection and mobility control
from the network side on how the different available network
Existing mobile terminals are generally equipped with accesses are used by the traffic generated by applications
multiple network interfaces, typically WiFi and cellular. This, running on the mobile.
together with the proliferation of femtocells, WiFi hot-spots,
and WiFi access to fixed residential home gateways, has IV. SDWN ARCHITECTURE
complicated the process of selecting the best access We next describe a generic software defined wireless
technology at each moment. A mobile node may decide to use networking (SDWN) architecture. Such an architecture aims at
the available attachment options sequentially (i.e., move all bringing the benefits of logical orchestration: by providing
traffic from one technology to another one) or simultaneously well-defined interfaces for control plane functions and
(i.e., move selected flows from one access to another [10]). enabling a richer flexibility in user plane traffic handling.
Despite that telecom operators can offer both residential fixed
and mobile services, WiFi hot-spot accesses are usually not A. System View
directly managed by the operator and their characteristics Figure 2 shows an SDWN-based architecture of a mobile
make it challenging to ensure a given QoS (as opposed to network operator, where a solid line in the figure denotes a
cellular accesses). Because of these factors, the decision of user plane connection, and a dashed one is used for the control
plane. We take the 3GPP Evolved Packet System as reference
architecture to link the proposed concepts with a well The brain of the architecture, the SDN controller, is connected
established and understood system architecture. to each programmable entity. Note that the SDN controller is a
A mobile network typically exhibits multiple heterogeneous logical entity, which might also be decentralized into different
Radio Access Networks (RANs) connected to a common physical boxes to improve scalability and performance, though
transport core network. Note the connection between the last this is currently the subject of extensive research [11].
network entity providing radio access and the core transport In order to allow third parties (such as service and application
network might involve a wired or wireless backhaul network providers) to influence/control the behavior of the network, an
(shown as part of the RAN in the figure) by using a API is enabled. This API effectively enables external players
combination of technologies (e.g., fiber optic, microwave) and to get access to the network resources, similar to what an OS
topologies (e.g., ring structure, daisy chain) in the backhaul does with the access of applications to computational
segment. Three well-know examples of RANs are shown in resources and peripherals. The API offered by the SDN
Figure 2: the UTRAN (for UMTS), E-UTRAN (LTE), and a controller supports different access levels to the external
WiFi hotspot. However, it should be noted that the proposed parties, so personalization can vary on different dimensions:
architecture is generic enough to support other RAN per application, per user, per (virtual) operator, per access
technologies as well, both already existing ones (such as network, or a combination of them.
WiMAX), or future ones.
B. Key Interfaces
In the SDWN architecture, radio access networks are
enhanced with programmability (as introduced in more detail We now focus on the description of the different interfaces
below), supporting multiple functionality levels to allow for (see Figure 3):
incremental deployments. The core transport is composed of • A northbound interface to the (virtual) operators
programmable L2 switches and L3 routers, allowing set-up of sharing the same physical set of network resources
unicast and multicast forwarding at the flow level (as allowing them to dynamically change the share of
supported for example by OpenFlow). Multiple (virtual) resources, for example to adapt to network load or to
operators might share part of the radio, backhaul and transport the number and profile of users attached to the
core network, which requires the interconnection of the core physical shared network at any given moment of
control plane entities – in charge of functions such as time. This interface should be able to implement
authentication, authorization, charging, subscriber richer SLAs as compared to the ones available
management, mobility management, QoS provisioning or nowadays, as a more dynamic and almost real-time
connection to external services/networks – with the reconfiguration of the network would be possible.
programmable network. Each (virtual) operator should have access to an
Two different models can be adopted to implement an SDWN abstracted view of its assigned resources, so they can
architecture: “evolutionary" and “clean slate.” The program that "virtual" network as a physical one.
evolutionary model allows for incremental deployment in • A northbound interface to the external parties
existing networks: legacy control plane entities from the (service and application providers) authorized to
operators can connect to the transport core network without influence the network behavior. This interface should
modifying the existing interfaces. In this model, the SDN be properly secured, granting access with different
controller implements standardized interfaces to support the granularities and permissions. The interface should
internetworking with existing legacy entities, even if they run be powerful enough to allow an application provider
on a virtualized environment (what is known as virtual EPC, to influence how its traffic is handled, even taking
vEPC). into consideration the virtual operator its users are
In the clean slate model, the control plane functions are getting access from. Note that this is possible because
directly programmed on the SDN controller or on top of it as of the centralization achieved by the use of the SDN
applications, using a software API between the virtual approach, though this may introduce scalability
operators and the SDN controller. While this approach does issues (e.g., up to per flow signaling, need for
not allow for an easy incremental deployment, it brings from frequent network monitoring, etc) that need to be
day one all the advantages of programmable network taken into account.
architectures. For example, the deployment of new network • A southbound interface to the physical user-plane
functions and services is much easier and faster, as it can be network entities in the core transport backbone. This
directly implemented on the controller and does not need to interface is used by the SDN controller to implement
impact on multiple interfaces and equipment from different the different behavior policies according to the
vendors. We can just take the simple, but very representative, requests from the external parties, the virtual
example of IPv6 support on a mobile network. With the clean operators associated to the different users attached to
slate approach, adding IPv6 support would just require the network, and the network conditions. Given the
additional code on the SDN controller, as compared to logical centralization provided by SDN, close to
defining new interfaces, procedures on the different control maximum utilization of the capacity of the network
and user plane entities, which require software/firmware links can be achieved. This interface also allows for
updates (if not even replacing some hardware). effective sharing of a common backbone and
 backhaul network by different operators, which may the different SLAs that the virtual operators may
even connect to the Internet via different gateways. have in place with their users.
• A southbound interface to the physical user-plane • A southbound interface with the mobile node. This
entities in the RAN. This interface allows for interface provides the network with certain
effective virtualization of the access network, programmability capabilities on the mobile node.
therefore sharing the same physical resources among This can be used for example to improve the mobility
different operators. Besides, this interface should experience, by better exploiting the simultaneous use
allow programming the wireless access technologies of available wireless access networks, e.g., helping in
to provide the expected behavior, depending on the access network and interface selection.
specific needs and characteristics of the mobile
terminal, the requests from the external providers and

Figure 3. SDN-based mobile network interface architecture

A proper implementation of the above interfaces, together bearers for 3GPP accesses, configuration of IEEE 802.11aa
with the required intelligence on the SDN controller will behavior for multicast transmission over WiFi, management of
provide new functions/flexibility, not available on today’s optical and microwave parameters in backhaul links, etc.
architectures. We next describe some examples. Overall, this programmability can be used to help meeting
The logical centralization of SDWN offers the ability to both global network-wide goals (in terms of efficiency), as
change the forwarding of user data traffic in the access (radio well as particular targets on a per virtual operator and service
and backhaul) and core transport networks, which can be used provider basis (e.g., differentiation on a per application/mobile
to provide functionalities such as mobility management and node).
QoS provisioning. The joint dynamic configuration of the RAN and transport
The programmable configuration of the RAN (including the core network allows performing global network updates to
backhaul), e.g., MAC and radio access parameters, allows for better adapt to current conditions. Examples of these updates
best dynamic use of available resources, considering current are using different unicast/multicast distribution schema
load, users' distribution and network sharing among virtual within the network in order to better meet specific QoS
operators. Examples of this are: configuration of the Wireless requirements. But this can also be used to offer dynamic
Multimedia Extensions for WiFi, configuration of radio adaptation of the traffic at transport and application level, for
example, NAT or IPv6/IPv4 transition functionality can be Extending the programmability to the mobile nodes enables
dynamically enabled and moved within the network. Even very interesting enhancements on the users' experience.
more, the forwarding path of selected multimedia flows can be Handover management can be made much more efficient, as
updated, so they traverse a middleware box capable of network and client side would be in tight coordination, which
adapting the content to the network and client conditions (by eases tasks such as cellular offloading and faster handovers.
re-encoding the multimedia stream, or selectively dropping Network discovery can also be simplified, as the network can
some packets, for example). enforce its policies more easily and almost in real time.

Figure 4. Case study: SDWN high-level interactions

this event can be reported to the SDN controller(s), using the

C. Case study
existing southbound interface. This attachment event is also
In order to show how the proposed SDWN framework works, notified to the Virtual Operator A, which can optionally
we next make use of an example case study, describing the trigger some specific configuration actions on the network for
high level interactions between the main SDWN components that particular UE. These additional configurations can affect
(see Figure 4). We not only explain how to make use of the both the transport core and the wireless access networks, for
defined interfaces – highlighting the opportunities enabled by example to offer a prioritized wireless access, by configuring
the use of SDN in a wireless mobile network – but also the MAC service provided to the UE.
identify some of the challenges posed by the adoption of such If the UE requests a service for which the network has an
an architecture. agreement in place (video in this example), its traffic is
In our case study – which is meant to serve just as an example, provided with a differentiated service. The network
and therefore it does not capture all the possible interactions – configuration required to do so can be already in place, or can
a virtual operator A operates using the resources of a core be triggered on the service provider by the SDN controller(s).
transport backbone and different radio access networks. Using Both approaches are possible (pre-configuration of the traffic
the defined northbound interface with the SDN controller(s), forwarding and on-demand dynamic configuration), allowing
Virtual Operator A is capable of pre-configuring the physical for different service models (e.g., gold users are provisioned
network resources, so it appears as a valid operator in the area on demand, while regular users are provided with a default
covered by wireless access networks. Similarly, a service service). Note that scalability might be a concern if pre-
provider can also pre-configure the network in order to configured policies are not used, and the traffic from all the
provide a certain default treatment to its traffic (e.g., provide users is treated on-demand. This is one of the main challenges
low end-to-end latency). Note that in both cases, the use of the to be addressed in SDN architectures.
interfaces provided to configure/influence the behavior of the If the UE gets into coverage of another wireless access
network requires proper service level agreements (SLAs) to be network, also managed by the same SDN controller(s), and
in place. with the required SLAs in place, this event is received by the
In this scenario, if a user equipment (UE) falls within the network, which can then evaluate to selectively move some
coverage of the Wireless RAN 1 and attaches to the network, flows to the new access, in order to improve this particular
UE’s QoE, the overall utilization of the network, or both. In parameters usage, which has specifically discussed the SDN
this example, video traffic is moved to the Wireless RAN 2, capabilities that are needed in the specific IEEE 802
by exploiting the existing SDN interfaces with both the technologies [13]; ongoing discussion within the IEEE 802.16
infrastructure network entities and the UE. WG, to introduce the bridging capability directly within the
wireless stack in such a way that SDN functionalities can be
V. STANDARDIZATION: STATUS AND FUTURE NEEDS applied to the wireless connections [14]; and the IEEE
802.11ak8 and 802.1Qbz9 specifications, which are very
A. Current efforts
relevant as they specify how 802.11 access points (APs)
Regarding standardization efforts in the area of SDN, the most interconnect with wired bridged technologies in a simpler (and
relevant Standards Developing Organization (SDO) is the more SDN-friendly) manner.
Open Networking Foundation (ONF)4, a member-driven The 3GPP is not currently addressing directly how to
standards organization aiming to promote and adopt Software introduce SDN concepts on their specifications although SDN-
Defined Networking through open standards development. based approaches are being investigated as a possible solution
ONF is the home of the well-known OpenFlow standard and to some current challenges. For example, the concept of
OpenFlow Switch Specification [5], defining the protocol used reconfigurable backhaul is used for RAN sharing. The sharing
for the communication between the OpenFlow controller and of the radio access network by different operators requires the
switches. In addition to these core standards, the ONF also traffic to be routed to the correct network, depending on the
publishes a testing specification to guide the conformance of operator policies. In addition to this functionality, SDN-related
the OpenFlow switches [12]. ONF is structured in several concepts are being discussed as a technology that can be
working groups (WGs), which address topics that range from applied to Self-Organizing Networks (SON) [15]. SON
extensibility (adding new features) to migration (for existing networks provide mechanisms for self-optimization, self-
networks to adapt to the OpenFlow standard). Recently, a healing and self-configuration, applying concepts such as the
Wireless and Mobile Working Group has been established to programmable control of the network, which are shared by
address the specific requirements of mobile networks. The SDN-based approaches.
charter of the group5 lists a number of identified use cases Other SDOs such as ETSI or ITU-T are also considering
ranging from mobile backhaul to mobile core issues. approaches similar to SDN for defining their architectures for
OpenFlow extensions can be expected in near future to cover Future Networks. This is the case of ITU-T SG 1310 (Future
the gap in existing specifications. networks including cloud computing, mobile and next-
At the IETF6, the SDN trend is impacting several WGs. The generation networks), which has included in its work
most representative SDN-related WGs are the following: the programme several topics related with SDN networks, as
Forwarding and Control Element Separation (FORCES), mechanisms for realizing the Network Intelligence Capability
which is tackling the separation between the control plane and Enhancement (NICE), developed at the ITU-T. In the case of
data plane; the Network Virtualization Overlays (NVO3), ETSI, the Autonomic network engineering for the self-
focusing on the data center overlay problem space and managing Future Internet (AFI) Industry Specification Groups
architecture; and finally, the Interface to the Routing System (ISG) is working on solutions for autonomic (self-)
(I2RS), which focuses on providing a real-time interface into management and control of the network resources in mobile
the IP routing system. networks.
In addition, the IRTF has created the Software-Defined Additionally, ETSI has recently (October 2013) delivered
Networking Research Group (SDNRG7), which is analyzing several specifications on Network Function Virtualization11
the approaches that can be used both in the near term and in (NFV). Although NFV and SDN concepts are mutually
the future. Finally, there are some other WGs at the IETF that beneficial but not dependent on each other, both proposals
can bring useful knowledge into SDN standardization, such as share the idea of providing a set of functions enabling the
ALTO, PCE, NETCONF, NETMOD and DMM WGs. programing of the network functions.
The IEEE 802 LAN/MAN Standards Committee has also
started SDN-related activities. Although there is currently no B. Future Needs
WG focused specifically on SDN in IEEE 802, there are Standardization efforts represent one of the key building
ongoing discussions on how to introduce SDN capabilities on blocks of the telecommunication industry. A standard answers
wireless and wired technologies. The most relevant activities to the need for communication among devices of different
are: the Open Mobile Network Interface for Omni-Range Area manufacturers. Hence the focus of a standard is placed on
Networks (OmniRAN) Study Group, defining an access protocol definition, message formats and corresponding
network specification based on IEEE 802 technologies (e.g., behaviors, leaving the specific implementations to the choice
802.11, 802.1X, etc.) that includes a network architecture, of the manufacturer. For example, OpenFlow, one of the most
recommended communication protocols, and link-specific relevant SDN tools available, uses a central controller and a
4
https://www.opennetworking.org/
5 8
https://www.opennetworking.org/images/stories/downloads/working- http://www.ieee802.org/11/Reports/tgak_update.htm
9
groups/charter-wireless-mobile.pdf http://www.ieee802.org/1/pages/802.1bz.html
6 10
http://www.ietf.org/ http://www.itu.int/en/ITU-T/studygroups/2013-2016/13/Pages/default.aspx
7 11
http://irtf.org/sdnrg http://www.etsi.org/technologies-clusters/technologies/nfv
set of interfaces, namely a Northbound interface to ACKNOWLEDGEMENTS
communicate with control applications and a Southbound The research leading to these results has been partly funded by
interface to communicate with the actual network hardware. the European Community's Seventh Framework Program
Through these two interfaces, applications can be developed to FP7/2007-2013 under grant agreement n° 317941 – project
use network functions and interact with the deployed iJOIN, http://www.ict-ijoin.eu/.
hardware. Hence, it is of vital importance to standardize these
two interfaces in order to be able to communicate and interact REFERENCES
with hardware provided by different vendors. [1] Y.-J. Chang et al., “Scalable and Elastic Telecommuncation Services in
Currently, different SDOs are working on very related and the Cloud”, Bell Labs Technical Journal, vol. 17, no. 2, pp. 81-96, 2012.
entangled issues although each one is focused on a different [2] “Network Functions Virtualisation – An Introduction, , Benefits,
point of view. As such, the current panorama of Enablers, Challenges, and Call for Action”, white paper available at
http://portal.etsi.org/NFV/NFV_White_Paper.pdf, 2012.
standardization activities is formed by a number of non-
[3] S. Jain et al., “B4: experience with a globallydeployed software defined
collaborative activities, several of them trying to solve the WAN,” in Proceedings of the ACM SIGCOMM 2013 conference on
same issues in a slightly different way. SDN relies on APIs SIGCOMM, ser. SIGCOMM ’13. New York, NY, USA: ACM, 2013,
and standardized Service Access Points (SAPs), to be able to pp. 3–14.
control the behavior of physical network elements. As such, [4] Open Network Foundation, “Software-Defined Networking: The New
standardization plays a crucial role for this purpose, since Norm for Networks”, 2012.
without a clear definition of these interfaces, SDN cannot [5] Open Network Foundation, OpenFlow Switch Specification, v1.4.0.
October 14, 2013.
become a reality. In order to organize and focus the work
[6] P. Calhoun, M. Montemurro, D. Stanley “Control And Provisioning of
performed by the different bodies addressing the important Wireless Access Points (CAPWAP) Protocol Specification”, RFC 5415,
labor of standardizing SDN related technologies, it is desirable March 2009.
to create a common view of use cases, network services and [7] G. Bianchi et al., “MAClets: active MAC protocols over hard-coded
functionality among all SDOs, which can be further used to devices,” in Proceedings of the 8th international conference on
Emerging networking experiments and technologies (CoNEXT '12),
create a real momentum and push forward the next generation ACM, New York, NY, USA, pp. 229-240, 2012.
of SDN concepts. [8] K. Pentikousis et al., "Mobileflow: Toward software-defined mobile
Table 1: The case for SDN in Mobile Networks networks," Communications Magazine, IEEE, vol.51, no.7, pp.44,53,
Key Benefits Key Challenges July 2013.
Easier deployment of new Specification of the interfaces [9] NGMN, “Backhaul Evolution - Integrated QoS Management”.
services [10] A. de la Oliva et al., "IP flow mobility: smart traffic offload for future
Reduced management and Need to integrate scheduled- wireless networks," Communications Magazine, IEEE, vol.49, no.10,
operational costs of based and contention-based pp.124,132, Oct. 2011.
heterogeneous technologies systems [11] A. Dixit et al. “Towards an elastic distributed SDN controller,” in
Efficient operation of multi- Harmonization of the Proceedings of the second ACM SIGCOMM workshop on Hot topics in
vendor infrastructures standardization efforts software defined networking (HotSDN '13). ACM, New York, NY,
USA, 7-12.
Increased accountability and Verifiable security and privacy
[12] Open Network Foundation, “Conformance Test Specification for
service differentiation architecture
OpenFlow Switch Specification”, v1.0.1, June 13, 2013.
Continuous and transparent Operation and management of
[13] R. Marks, A. de la Oliva and J.C. Zuñiga, “Proposed OmniRAN SDN
enhancement of network wireless networks is more Use Case for External Communication”, August 2013.
operation complex
[14] IEEE P802.16r, “View of Connection-Oriented Software-Defined
Networking for Wireless Backhaul of Small Cells”, July 2013.
VI. CONCLUSIONS
[15] 3GPP, TS 32.500, Telecommunication Management; “Self-Organizing
In this article we have identified the opportunities that Networks (SON); Concepts and requirements”, December 2011.
software defined networking can bring to wireless and mobile
networks. We propose a high-level architecture leveraging on BIOGRAPHIES
the advantages of the logical centralization provided by SDN. Carlos J. Bernardos received a Telecommunication
We have first defined the main functions that should be Engineering degree in 2003, and a PhD in Telematics in
2006, both from Universidad Carlos III de Madrid (UC3M),
supported by a mobile SDN architecture, and then specified where he worked as a research and teaching assistant from
the required interfaces and describe some of the interactions 2003 to 2008 and, since then, has worked as an Associate
that would be needed to enable new and/or richer use cases. A Professor. His current work focuses on mobility in
summary of the key benefits and challenges for SDN in heterogeneous wireless networks. He has published over 50
scientific papers in international journals and conferences, and he is an active
mobile networks is provided in Table 1. contributor to the IETF. He has served as guest editor of IEEE Network.
Last, but not least, we have reviewed ongoing standardization
efforts around SDN topics, identifying the future needs to Antonio de la Oliva received a Telecommunication
ensure a successful practical deployment of SDN mechanisms Engineering degree in 2004, and a PhD in Telematics in 2008
in the wireless arena. (receiving a national award ex-aequo for the best PhD. thesis
in IPTV services) , both from the University Carlos III of
Madrid (UC3M), where he worked as a research and teaching
assistant from 2005 to 2008 and, since then, has worked as an
Visiting Professor. His research is focused to mobility in
heterogeneous networks and wireless systems. He has
published over 30 scientific papers in prestigious international journals and Hao Jin is a senior engineer at InterDigital
conferences, and he is also an active contributor and voting member of the Communications, Inc., where he is currently working on the
IEEE 802.21 where he has served as Vice-chair of IEEE 802.21b and research and development of SDN and network
Technical Editor of IEEE 802.21d. Currently he is involved in the FP7 virtualization for wireless and mobile networks. He earned
CROWD project. his B.S. degree in Electrical Engineering from Nanjing
University, China in 2006, and his Ph.D. degree in
Electrical Engineering from Florida International
Pablo Serrano got his Telecommunication University in 2012. His research interests include SDN,
Engineering degree and his PhD from the network virtualization, next generation wireless networks, and data center
Universidad Carlos III de Madrid (UC3M) in networking.
2002 and 2006, respectively. He has been with
the Telematics Department of UC3M since
2002, where he currently holds the position of Luis M. Contreras completed a six-year Telecom Engineer
Associate Professor. He was a visiting degree at the Universidad Politécnica of Madrid (1997),
researcher at the Computer Network Research Group at Univ. of and holds an M. Sc. on Telematics from UC3M (2010). In
Massachusetts Amherst in 2007, and at Telefonica Research Center in 1997 he joined Alcatel in Spain taking several positions in
Barcelona in 2013. He has over 50 scientific papers in peer-reviewed both wireless and fixed network fields. In 2006 he joined
international journal and conferences. He serves on the Editorial Board of the Network Planning department of Orange in Spain
IEEE Communications Letters, and has been a guest editor for Computer (France Télécom group) taking responsibilities on the IP
Networks. backbone and mobile packet core planning. Between 2002
and 2010 he was also adjunct lecturer at the Telematics
department of UC3M, where he is currently a Ph.D. student. Since August
Albert Banchs received his degree in telecommunications
2011 he is part of Telefónica I+D / Telefónica Global CTO, working on SDN,
engineering from the Polytechnic University of Catalonia in
transport networks and their interaction with cloud and distributed services.
1997, and his Ph.D. degree from the same university in He is active contributor to IETF and ONF.
2002. He received a national award for the best Ph.D. thesis
on broadband networks. He was a visiting researcher at
ICSI, Berkeley, in 1997, worked for Telefonica I+D, Spain, Juan Carlos Zúñiga received his engineering degree from
in 1998, and for NEC Europe Ltd., Germany, from 1998 to the UNAM, Mexico, and his MSc and DIC from the
2003. He has been with the University Carlos III of Madrid Imperial College London, UK. He is a Principal Engineer
since 2003. Since 2009, he also has a double affiliation as Deputy Director of at InterDigital, currently leading the standardization
the IMDEA Networks institute. activities in the areas of heterogeneous networks, software
defined networking, distributed content management and
Internet privacy. He has actively contributed and held
leadership roles in different standards fora, such as IEEE
802, IETF, and 3GPP. He has been with InterDigital since
2001. Previously he worked with Harris Communications in Canada, Nortel
Networks in the UK, and Kb/Tel in Mexico. He is an inventor of more than 30
granted U.S. patents and 70 published applications.