|Issue:||Latin America 2015|
|Topic:||What are the reasons behind software-centric networks?|
|Title:||Principal Analyst LATAM|
Arivaldo Lopes is a Principal Analyst with Ovum for Latin America. He leads Informa’s efforts in understanding the Latin American operators’ strategic options for investment in 4G technologies, as well as providing competitive benchmark analysis, country profiles and analyzing the opportunities for MVNOs in the region. His research also focuses on opportunities in the mobile Internet market for Latin American operators, developers and media companies, given the growing polarization on two or three mobile OS ecosystems.
Arivaldo joined Informa in April 2012. Prior to that he was a Senior Financial Analyst with Google, where he was in charge of capacity planning and forecast for the SMB Customer Services organization. Arivaldo also has eight years’ experience in the telecoms market, having worked at Vivo as a CRM analyst, and as pricing manager at Brasil Telecom and Oi.
Arivaldo holds a Masters degree in Communications Management at the University of Strathclyde, and has a degree in Economics from the University of Campinas.
The erosion of traditional revenues, especially voice, in telecommunications networks, due to the growing data traffic, has put pressure on telecoms operators, from one side they have to find new revenue sources to replace old ones, and on the other hand, the way legacy networks are managed have been found to be inadequate. SDN and NFV are technologies that are part of the transformation that these operators are experiencing and this article reviews definitions, challenges and opportunities around software-centric networks.
The fast development of the internet, and the consequent increase in data traffic has created enormous pressure on telecommunications networks. Even regions like Latin America, that traditionally is slow in to adopt the same trends seen in more developed areas, demand for broadband has been high both on mobile and fixed networks, many operators in the region, for instance, are already reporting that smartphones represent more than 80% of device sales.
Telecom operators are at the beginning of a process that will see them transform their business and operations in order to cope with this change in demand. Software-defined network and virtualization are part of this transformation in the network side, they will change how networks are planned, built and operated. These network changes will enable a more flexible, scalable and open network.
In terms of definition, software-defined networks (SDN), refers to an emerging architecture that allows for the abstraction of the control and data planes in a network. The separation of these two functions brings, among other benefits, the possibility to directly program the network control independently of the underlying infrastructure. A central piece in SDN architecture is OpenFlow, the communications protocol between the control plane and the infrastructure, which enables the abstraction between these two layers. According to Open Networking Foundation (ONF), an organization that develops OpenFlow, the key trends that are driving the need for new network architecture are:
• Changing traffic patterns – geographically distributed databases hosted in public and private clouds required flexible traffic management.
• The “consumerization of IT” – widespread use of employee’s devices for business purposes means networks are required to be more secured than before.
• The popularization of cloud services – enterprises and consumers are using services on the cloud, requiring on-demand access to applications.
• Big Data – analysis of massive datasets require parallel processing, ever higher demand for capacity.
The other key concept frequently associated with SDN, even though, it is not confounded with it, and it is Network Functions Virtualization (NFV), at its core NFV is the use of IT virtualization techniques to network components. It proposes to separate hardware from software, implementing network elements in the data centre, and, in the process, moving away from proprietary to commoditized hardware, on other words, it envisages a fundamental change in the way telecommunications infrastructure market is structured, around proprietary solutions. A key selling point of NFV is the promised agility in developing new services without the need to add more purpose-specific equipment consuming floor space and power.
As mentioned before, SDN and NFV are complementary technologies, while the first is focused on network programmability, the second is on resource flexibility in the network core. What is expected from this move towards IT is that this will bring flexibility and scalability to network design, based on open standards. Overall, the main driver is cost savings in the network, as it is recognized that legacy networks are ill-suited to cost-effectively scale capacity, however, it is expected that other benefits will follow, especially it will open the network for in-house or third-party innovation.
Why the interest in SDNs?
Even though these are early stages of software-centric networks, the change towards a merge of telecoms and IT world is inevitable, the business case for the disruption is strong for a number of reasons, among them: network programmability allows for rapid innovation; the combination of commodity hardware, open standard protocols and software-based networks means there are an opportunity for breaking vendor lock-in in the infrastructure market; the use of common hardware also envisages cost reduction; aligns telecoms with the rapid pace in computing power; a key issue for operators is the non-alignment between revenue growth and traffic demand among telecoms operators, which requires more efficient ways to run the network; support is growing among key players in the industry, from telecom operators, vendors, standards organizations and academic institutions.
Given all the hype with SDN and NFV is easy to lose sight of the real objectives in introducing these architectures, the technology per se is not the goal, they are a step towards modernizing the way telecoms networks are procured, designed, integrated, deployed and maintained. Ultimately, the real objectives are scalability, flexibility, automation, programmability, and in the end, cost savings and fast deployment of new services.
The problem with traditional networks is that the time required to develop new services is too long for a market that is demanding fast innovation. Additionally, SDNs can bring automation to the control and provisioning processes, the efficiency derived from this has the potential to reduce OPEX.
Latin American operators of all sizes are already involved in testing and eventually, launching, services due to SDN. A good example of new opportunities made possible by SDN are on-demand services like QoS or bandwidth per application and per user, even considering the purchasing process made available to customers on an online portal. This is the case of Mexican operator Marcatel, which offers on-demand bandwidth which customers can acquire on a self-service basis in an online portal. However, the main developments, especially in NFV, are occurring on test labs, ETSI, the entity working on developing specifications for the technology, counts more than 18 PoCs already completed and another 16 on-going. Nonetheless, implementations are following a pattern, of incremental adoption, based on domain specific solutions with clear business or operational goals, i.e., new service launch, cost reduction or operational efficiencies.
Software-centric networks are the most important trend in terms of network modernization, among the problems with legacy networks that led to the research of SDN and NFV for telecommunications, there is the overprovisioning of capacity in order to deal with peak demand, however, it is no longer a sustainable business model anymore.
On the other hand, operators need to aware that adoption of SDN/NFV will happen gradually, on a piece meal basis. Standards are still under development, different implementations, albeit based on open standards, not necessarily means interoperable and not necessarily means free of vendor lock-in. There are still a lot of work to be done in order to develop truly open standards that can fulfil the promises of SDN/NFV.