Background
To understand what SDN is and how it applies to the modern network infrastructure it’s good to highlight the past design and most importantly the most used network designs seen today. The traditional approach of networking will have operations and functionalities implemented in specific devices. Therefore, a switch or a router will have specific instructions, in essence, this creates dedicated appliances that are only responsible for what is assigned to them. Moreover, these functionalities are also implemented using dedicated physical devices (hardware) such as application specific integrated circuit (ASIC). However, SDN offers another perspective where it promotes a virtualized network architecture that bargains on software defined systems.
In essence, SDN is a technique that approaches network design in terms of abstract and separate elements. Therefore, the processing of functionalities does not take place within the same device, instead, it decouples the control information from the data itself. The end results is a control system (application) that governs the network through intelligent policies made by network experts. Furthermore, this solution is dependent on the foundational concepts of OpenFlow where forwarding access is given to network devices. OpenFlow is a communication protocol that grants access to network devices such as routers and switches to the forwarding plane of any given network. A consequence of this functionality is that researchers and network engineers can run experiments on new design protocols without exposing the structures of network devices. Hence, OpenFlow is just but a subset of the broad concepts of SDN.
Fig: SDN Design Approach
In general, SDN promises to offer a more centralised and simplified control of networks, therefore, large organisations will have the convenience of managing a network in a manner that is close to an information system such as an enterprise system. Furthermore, depending on the vendor of the SDN application, these benefits will increase to cater for user requirements. Consider an SDN application such as Microsoft Hyper-V network virtualization that offers a free operating system. Unlike other applications such as VMware NSX, this application does not require a licence to use, instead, the entire application is bought as a holistic package which lowers the costs on top of the other SDN benefits.
SDN Critical Evaluation
Nevertheless, what are the technical or functional benefits of SDN in relation to the Microsoft Hyper-V application? First, is a centralised control of networking devices regardless of the vendor. Network administration is an integral part of an organisation that ensure business activities such as communication are done with minimal interruptions. Traditional networks make it difficult to manage a network and its components. For instance, an administrator is forced to manage a physical switch or router based on its configurations. This problem is made worse when more than one device requires management. However, with this application, the administrator is provided with a management interface (Application program interface API) within the SDN console itself.
In addition to this, the management is able to have a reduced overhead for network control where user items are virtually isolated without the need for physical components. Consider an organisation that has various departments, to isolate this section VLANs (virtual local area networks) are created, however with Hyper-V, isolation is done using SDN. Furthermore, the system collaborates the needs of the virtual world with that of the physical environment. For instance, network modification requires physical adjustments collaborated with configurations. However, with Microsoft Hyper-V, a single control that is managed by a single individual can perform this task without the collaboration of various teams [5]. Moreover, consider the time saved by the virtualization of a given network architecture. Network requirements such as upgrades are done conveniently to all devices at a go. Microsoft hyper-v SDN application also has snapshotting capabilities which make it easier to recover from failures since the configurations are stored within its database system.
Finally, SDN offers isolation and secure traffic control through access control systems as well as through firewalls. In essence, service providers can offer a centralised control of their processes using the management APIs. Moreover, the network administrators can define traffic policies and rules using the management console which enhances traffic control. Furthermore, since SDN applications such as Microsoft hyper-V are software in nature, it’s easier to provide upgrades and capability extensions based on the network behaviour and requirements.
An overall characteristic that would define software-defined networks is a disruptive force, a disruptive force that is both innovative and necessary for all modern organisations. Moreover, its existence affects all major players within the industry from vendors to network operators. However, several critical enhancement has been seen and will lead to other future network advancements. Among them is programmability of a network where, both the control structure and the data plane are conveniently defined. Therefore, the deployment of network functionalities is made faster speeding up the innovation process of future hardware and software components.
Potential of Network advancements
Virtualization also is set to break the barriers observed with wireless networks where again providers are limited by their budgets in providing services that meet the current user demands. When SDN is completely incorporated into network designs, the overall implementation cost of such services will be lowered facilitating better wireless solutions that are met with the necessary equipment’s for instance SDN enabled switches for wireless networks. Finally, network openness is a potential network advancement that will be characterised by SDN networks. In it, SDN provides open standards that can be used to develop an open minded society that is free to think thus improving the speed of innovation. Consider OpenFlow for example, an open standard that facilitates research of network systems through fast testing and prototyping.
Modern network designs require highly skilled personnel to meet the installation and configuration requirements. These individuals such as network administrators manage the complex interaction between network nodes such as those of a router and a switch. This approach is difficult, to say the least as the existing systems are designed to have isolated control i.e. the programming interfaces are meant for a specific device having them. Therefore, even with the desire to transition into a system-based design, the existing infrastructure acts as a barrier to unifying network administration. Furthermore, many of these devices are supplied by multiple vendors which increase the operational costs. A network within a given organisation will have network devices and applications from multiple vendors who use different management systems. Moreover, the administration of a network is made difficult by the existing network technologies.
Now, consider this, a network administrator hoping to transition into an SDN approach will have to deal with multiple vendor equipments’s which are operating within a diverse range of technologies. In addition to this, the current financial trends have reduced the operation costs of implementing network systems. This trend is facilitated by the scarcity of resources that has raised the standards of living thus necessitating a reduction of the non-basic requirements. However, despite the challenges, the solutions to both the network designs and scarcity of resources lies in a unified networking management system such as SDN. Finally, when implemented, SDN proponents are faced with a scalability challenge, where virtualization will have increased data requirements. An SDN application like the one discussed above (Microsoft Hyper-V), relies on a central controller i.e. the SDN console. Since it’s a single device, it’s not easily scalable moreover, in case it fails the entire network is deemed irrelevant thus creating a single point failure [9].
Benefits of SDN
According to Morton (2016), organisations should evaluate their existing network infrastructure before embarking on a complete SDN transformation. To date, SDN still faces considerable challenges most of which will be fixed with time and research. However, a critical evaluation of the implications of SDN is needed before implementation. For instance, for a wireless network, it’s appropriate to check whether the existing modules such as access points meet the minimal objectives for an SDN system. Furthermore, additional management tools are needed to integrate an SDN architecture, their purchase and support cost must be considered before developing and an implementation plan. Through these careful evaluations, the overall cost of establishing an SDN network is maintained to a minimum.
Although basic design standard exists for network devices, these equipment lack some of the necessary nodes to link to SDN consoles. For instance, an SDN-enabled switch was identified above for a wireless network, this because the current switches are designed as standalone devices that are controlled by the existing configurations. In the future, the industry’s stakeholder should develop standards that unify the design to have the basic requirements for SDN systems. Moreover, the standards developed should have a scalable component that minds the future outcomes similar to what is seen in IPsec protocols [10]. These suggested solutions provide a critical evaluation of the existing problems by providing answers to both the immediate and future problems.
Through its abstractive design, SDN can create a granular control of mobile networks that are changing at a fast and drastic rate. Mobile networks today are characterised by wireless technologies such as 4G, LTE and Wi-Fi that are designed to accommodate the increased demand of data traffic. Most of the literature and research done so far focuses on the convergence of wireless networks in order to benefit from SDN components. This outlook is completely justified by the current proceedings that outline wireless networks as the most popular means of accessing the internet. Furthermore, this outcome comes with an increased demand for resources that require a scalable design in service delivery.
These requirements are easily met using an SDN approach that separates the network infrastructure from the control applications. Three major elements are identified by this approach; the end user (mobile gadgets), wireless mobile network (WMN) and the control plane. At the WMN section, SDN-enabled devices are implemented, these devices hold the network data thus are also identified as the data plane. On top of this section is the control plane where SDN APIs are used by network managers to offer control. This approach is convenient as it offers dynamic solutions for network demands particularly when upgrades are needed.
Challenges with implementing SDN
Fig: Wireless
However, two areas of application come short when considering the implementation of SDN designs in wireless networks, those of security and standard requirements. Although standard wireless security procedure exists (IPsec, WPA2, and AES etc.) the existing literature fails to highlight specific protocols that solely focuses on SDN networks and the possible security vulnerabilities that may occur. For one, this approach will integrate devices from different vendors and some protocols may be obsolete in other creating access loopholes. Regardless of the outcome (secure or not) thorough research is needed to alleviate the doubt on the security of the systems. In addition to this, little is mentioned in the proposed standards of SDN products. A critical challenge of SDN implementation is that of interface requirements where some devices lack the necessary node to integrate to SDN consoles. Accurate proposals are needed for future equipment so that they meet the minimal integration requirements.
In an idealistic world, SDN would have the best features and functionalities promoting a unified network architecture, however, based on the current trends that include implementation challenges SDN is set for a problematic future. For one, it’s an open source technology which exposes it to many security violations. Therefore, to sort this security problem as stated above research is needed. Furthermore, a narrow perception is associated with SDN where the industrial stakeholders are only concerned on how it will either improve or hurt their networks. However, a broader perception is needed to make it palatable and suitable for the future environment. In addition to this, this technology is human centred, unlike other networking technologies that are device centred. In a nutshell, it will be difficult to convince managers and organisations to adopt a design that can be completely compromised by a single individual .
Nevertheless, every new technology faces its share of challenges and with time gets to overcome them. Similarly, SDN must adapt to the existing environment to offer the benefits it has but in a way that meets the needs of the users. These requirements include both technical and policy/rule needs. For instance, in the future SDN must address the security concern identified above. Moreover, it must be able to anticipate future network requirements and thus provide pre-determined solutions. A good example of this is scalability, with the current designs the central control may lack the necessary storage requirements and therefore lower its practicality.
Solutions for SDN challenges
Conclusion
A critical review of Software Defined Networks (SDN) identifies it more so as a technological aspiration rather than a technology in itself. In essence, the aim is to separate the data from the control features, therefore, have two broad categories of any given network i.e. the data plane and control plane. Through this process, networks will become programmable which will make implementation easier and faster. Moreover, through its open source policies, it will eliminate the monotony held by certain vendors in developing networking equipment. Finally, it will facilitate the benefits of virtualization where control, flexibility, efficiency and even scalability are enhanced. Consider the application identified above Microsoft hyper-V, its integration to the cloud (a crucial future technology) is easy, unlike traditional networking systems.
Nevertheless, despite its wide range of benefits it also does has its associated limitations and challenges. However, most of these challenges and limitations can be easily solved using research as seen in other technologies. To date, its implementation is limited to a few companies but even then it’s still seen as a proof of trial project where testing are still been done. Therefore, to a have a widespread application, SDN needs to prove its worth in these organisations by offering the proposed solutions and based on the current results, this outcome will surely be achieved.
References
Citrix. (2017). SDN 101: An introduction to software-defined networking. Netscaler ADC. [Online]. Available FTP: https://www.citrix.com/products/netscaler-adc/resources/sdn-101.html
Cisco. (2013). Software-Defined Networking: Why We Like It and How We Are Building On It. White paper. [Online]. Available FTP: https://www.cisco.com/c/dam/en_us/solutions/industries/docs/gov/cis13090_sdn_sled_white_paper.pdf
Sharma. N. (2015). Eight Big Benefits of Software-Defined Networking. [Online]. Available FTP: https://www.serverwatch.com/server-tutorials/eight-big-benefits-of-software-defined-networking.html
Rouse. M. (2017). Software-defined networking (SDN). Tech target. [Online]. Available FTP: https://searchsdn.techtarget.com/definition/software-defined-networking-SDN
OCEDO. (2016). Traditional vs Software Defined Networking. IP knowledge. [Online]. Available FTP: www.ipknowledge.net/wp-content/uploads/2014/12/SDN.pdf
Calsoft labs. (2016). Magic of SDN in networking. [Online]. Available FTP: https://sdn.calsoftlabs.com/downloads/WhitePapers/Magic-of-SDN-in-Networking.pdf
Ding. A, Crowcroft. J, Tarkoma. S & Flinck. H. (2014). Software Defined Networking for Security Enhancement in Wireless Mobile Networks. [Online]. Available FTP: https://www.cs.helsinki.fi/u/yding/pre-print-comnet2014.pdf
Comsoc. (2013). Are we ready for SDN? Implementation challenges for software-defined networks. IEEE Comsoc technology news. [Online]. Available FTP: https://www.comsoc.org/ctn/are-we-ready-sdn-implementation-challenges-software-defined-networks
Morton. D. (2016). Challenges of Software-Defined Networking. YDT. [Online]. Available FTP: https://yourdailytech.com/networking/challenges-of-software-defined-networking/
Nilsson. B. (2015). Migration to SDN Can Be Simple: A Case Study. Extreme Networks. [Online]. Available FTP: https://content.extremenetworks.com/h/i/319065573-migration-to-sdn-can-be-simple-a-case-study
O’Neill. T. (2015). What is the future of SDN? TAP into technology. [Online]. Available FTP: https://www.garlandtechnology.com/blog/what-is-the-future-of-sdn
Open networking foundation. (2017). Software-Defined Networking (SDN) Definition. [Online]. Available FTP: https://www.opennetworking.org/sdn-resources/sdn-definition
Open networking foundation. Software-Defined Networking: The New Norm for Networks. ONF white paper. [Online]. Available FTP: https://www.opennetworking.org/images/stories/downloads/sdn-resources/white-papers/wp-sdn-newnorm.pdf