|Issue:||Europe II 2014|
|Topic:||The art of self-disruption: How to transform your business with SDN|
|Title:||Head of EMEA, SP Architectures & CTO|
Paolo Campoli is Head of EMEAR SP Architectures & CTO in Cisco. His responsibilities include overlay sales prime of IP Routing, Optical, OSS and Service Provider Datacenter, where he leads network and process transformation projects for key accounts in EMEAR. Paolo is also the spokesperson of Cisco at key Industry events and with Regulatory Bodies. He joined Cisco Systems in 1998, as EMEA Consulting Engineer, involved in architecture design of DOCSIS, Broadband Access and Mobility solutions with major European carriers. From 2001 as Senior Manager of the Central Marketing organization of Cisco Europe, Paolo led development and packaging of Vertical industry solutions for the Enterprise market as well as triple play and Fibre-to-the-Home (FTTH)solutions for Service Providers. In 2004 he moved to Sales as Director of SP Business Development and was also appointed to lead Service Providers CTO interaction, SP architectures and technical operations. In 2010 the Consumer Video vertical unit, resulting from the acquisition of Scientific Atlanta and Kiss, was moved to report to Paolo, to create the Cisco’s first integrated business and technical team linking Network, Devices and Cloud. In 2011 his responsibilities were expanded to cover both Europe and Emerging Markets, as head of the SP Architectural play and regional CTO for the EMEAR Service Provider region.
Previously, from 1992 to 1998 Paolo was at Siemens-Italtel, authoring ETSI technical standards, heading the ATM project office and leading market development for the DVB Cable and Satellite digital Set Top Boxes.
Paolo graduated in 1990 with full marks and honours (100/100 cum laude) as Doctor in Electronics and Telecommunications, at the University “Politecnico of Milan” where he also developed one of the first comprehensive mathematical modeling of ATM network traffic, under mentorship of IEEE fellow Professor Decina.
Is SDN the new Holy Grail? It is shrouded in myths. The white label myth is that great savings on hardware will be realized, but in fact CAPEX savings are far lower than OPEX with SDN. The myth that bandwidth is expensive no longer holds true, where the economy-of-scale has changed things. The main benefits are not there. They are in the modularization of operations, for agile assembly of new services with fulfilment and service assurance capabilities, so that ‘fast failure’ can be tolerated, and spiky success can be exploited, as for internet applications. No two operators will make the journey towards SDN. They may approach it as a green-field expansion deployment or migrate to it on a service-by-service basis.
There are very few of us who possess the natural ability to challenge the norm, to look at a problem as an opportunity to innovate, or to disrupt what seems to work, or at least isn’t broken. The fact is even more evident in our industry, where there is a lot of organization, process and infrastructure legacy. It is an industry where there are many customers who are reliant on an ‘always on’ experience, and more worrying where a dropped ball can result in consumer outrage expressed over social networks, which can have a direct impact on the bottom line.
The telecom industry has been going through several major transitions – cloud, mobility, applications, big data, internet of things – all of these market disruptions have put major stresses on the infrastructure and the business models, as well as having a seemingly irreversibly eroding effect on revenues. Now, there may be a holy grail within reach, something that will ensure the sustainable future of the industry. If only we can quantify the actual benefits and mitigate the risks that could result in the self-disrupting effects of SDN.
SDN – Hype or Holy Grail?
There are a number of assumptions in telecoms that stop us from innovating in the way we need to. They are in most cases quite easy to address.
The white label router myth: SDN promised cheaper networking hardware. By separating the control and data planes, you can run a network using commodity, white label hardware. A nice reduction in CAPEX, but is that too short-sighted? When we started looking into what having a programmable network means for operators, and combine it with virtualization of the network and network functions, we discovered that the real benefits of SDN comes from other areas:
– Firstly, being able to automate what were very manual, laborious operations. When you consider that 80% or more of a telco’s spend is Opex (based on Gartner’s study of traditional service provider expenditure, 2013), then a 10% reduction in Opex would have an 8% impact on costs, compared to a 10% reduction in Capex, which would only deliver a 2% reduction.
– Secondly, network capacity that is elastic and can flex according to loading demands. The key to creating a software-defined network with highly automated capabilities is to have a data plane that appears to have infinite capacity. We know this isn’t achievable, but having an elastic network fits the bill. Which leads us to exposing the second myth.
The expensive bandwidth myth: Network operators are focused on bandwidth optimization and traffic management. Bandwidth has always been an expensive commodity, but not anymore. Today, it’s cheaper to deploy a 100 Gbit/s link than it is to maintain a 10Gbit/s one or to upgrade to 40G. Even more so when Optical WDM is integrated in IP routers New technologies drive improved economies of scale and radically better economics.
There has been a lot of hype around software-defined networking (SDN) and Network Function Virtualization (NFV). However, now that we are working with operators on real use cases, the potential benefits for the industry to be more innovative, increase efficiencies and reduce operational expenses are much more tangible.
Disrupting business models
The legacy way of creating and launching services required time, energy and money which did not create an environment for innovation, where fast failures are part of the process. Currently, for a new service to make a return on investment, the service requires 30-40% adoption rates.
For telcos to compete at the pace of change that the broader technology industry is capable of, agility, ease of experimentation and fast failure are needed to make it possible to innovate at speeds that pure Internet players do. To do so, the architecture of their networks needs to be very different.
The ideal scenario is one in which telcos can provision, launch and test new application or services in a matter of days or weeks (in comparison to months or years at present) in a highly automated way, using a set of software building blocks with far less strain on resources. If the services stick – great, if they don’t – the investment was minimal and they can move on and try something else. Success also starts to look different. The margin to profitability becomes much smaller with profit realized with only 10% of subscribers taking up a new service.
The operator’s business justification for adopting SDN and NFV now takes into account an expanded set of KPIs (Key Performance Indicators). No longer is it just the reduction in CAPEX, we can add OPEX optimization, speed of access to new revenues and new markets and creation of new business models such as pay on demand. This is the ultimate benefit of SDN and virtualization. When the focus shifts to the applications and services and what they need from the network, not whether the network is built with enough capacity to support them. It is at this point that the telco becomes a real enabler.
Setting the network up for success
An operator’s network is a large, complex system, but we can simplify our view of the system by describing it in three layers. At the base is the programmable infrastructure layer, which comprises the computing, storage and network functions that can exist as physical and virtual elements. The applications layer sits at the top, and the management and orchestration layer in the middle. This middle layer is effectively a powerful software platform that simplifies and automates the control and orchestration of the network infrastructure and resources across all domains – wired, wireless, cloud and video – and across all layers.
Modularization of operations, a practice that relies on the notion of creating modules that can be reassembled for quicker service creation in the fulfilment and assurance domains, is an integral part of the new network set-up. There is limited benefit to NFV if operations are not modularized, so these things are two sides of the same coin.
Mastering the art of self-disruption
We’re already seeing the application of SDN and NFV in our customers’ networks. Some are lab set-ups like the JOLNET at Telecom Italia, but there are also live production networks such as Hrvatski Telekom in Croatia. We can see the benefits from a real use case with the deployment of a cloud video service for Hrvatski Telekom in only 50 days. Based on the approach taken by Hrvatski Telekom to build an all IPv6 software defined network with virtualized functionality, they were able to utilize functions that already existed and apply a logic to create an orchestrated workflow that greatly reduced time to market and reduced the expense that would usually be associated with building such a service.
How does the operator get their infrastructure from the current state towards this new software-defined state, while keeping all their existing customers happy? If you are thinking about redesigning an airplane while it is still in flight, it’s not going to happen.
No two operators will make the SDN journey in the same way. One may approach it from the network up as a green-field deployment, as Hrvatski Telekom has done. Another will do it from the service down, migrating on a service-by-service basis.
Operators will have to strike a balance between the business benefits and the exposure to risk of transforming their network in such a profound way. Unlike the quest for the holy grail, the strive by operators to be more agile, more innovative and to return to a sustainable business model is now very tangible.