McKinsey & Company, an American management consulting firm, has noted that between 2007 and 2018, the telecom industry in Europe and North America has experienced a decline of six and seven percentage-points respectively in the EBITDA margin. During this period, North American telecom revenue grew 20%, 28% points less than the North American market. In Europe, the picture is even bleaker, with a 24% decline in revenue over the same period compared with market growth of 18%. (McKinsey & Company: Telecom operators: Surviving and thriving through the next downturn – August 2019).
At the same time, investments in CSPs’ networks continue to grow to meet customer demand for capacity and new services like 5G. CSPs’ rising debt and capital intensity – a capital effectiveness indicator that measures capital investments over revenues, which is high compared to other verticals – between 10% and 20%, force CSPs (e.g. AT&T and Telefonica) to limit their capital investments by using infrastructure sharing, carving out some of their businesses (like mobile towers), or by completely disinvesting some parts of their business.
Poor Cost Structure of the Legacy Networking Pricing Model
The main problem seems to be how switches and routers are traditionally sold and the supplier relationship behind it. The legacy chassis-based pricing model is based on a perpetual license for hardware components including chassis, route processors and capacity-based line cards (calculated in number of 100G or 400G ports) while software is sold as a perpetual or time-based license per port. The issue is that CSPs have multiple networks, each of them built upon dedicated proprietary hardware usually provided by two different vendors. The bottom line is lots of different proprietary network-specific hardware whose cost grows linearly with traffic demand, leading to a poor cost structure. This model also encourages vendor lock-in, which has raised complaints about slow technological improvements, expensive upgrades and high maintenance fees. Some CSPs have even complained that due to their strong position in many long-term accounts, big traditional vendors have benefitted from high margins while they continue to struggle with high capital intensity and margin squeezes. Is there a way for CSPs to improve their bottom lines? How can CSPs benefit from the disaggregation business model… and make their CFO happy? Disaggregated networking solutions are not only changing the network architecture that has been in place for the past 25 years but are also enabling CSPs to reduce their capital use, improve their margins and cash flows – and have a more viable business model with their suppliers. At DriveNets, we have extensively discussed business and TCO models with our customers and the conclusion is that the disaggregated network business model can help CFOs in many different ways.
Network Disaggregation Lowers CapEx Per Port
Most CSPs are closely looking at their cost per bit – or cost per port when it comes to routing – in order to monitor and improve their margins. The general trend is that cost per bit is going down due to better procurement power of CSPs (like Buyin’s procurement alliance between Orange and DT) and technology improvements (higher port density). But these are not sufficient to compensate for the continued downfall of revenue per bit. The network disaggregation model offers a solution. It turns networks from hardware to software-centric, leading to simpler and standardized hardware across all network services (core, aggregation, edge) with a modular networking software sold independently of the hardware. On the hardware side, this pushes costs down thanks to the healthy competition between multiple vendors and economies of scale across all network services. And when multiple network services share the same physical location (“network colocation”), they can share the same hardware, leading to less hardware overall and bigger savings. On the software side, CSPs can pay only for the features they need, either on a time-based, perpetual or subscription license (OpEx expense). Overall, the CapEx per bit is expected to be much lower than it is today.
Network Disaggregation Lowers OpEx Per Port
Network standardization and simplification across multiple networks push fixed and variable operational expenses down. The unified software reduces the fixed security and upgrade efforts across the entire network. The cost of installation, service assurance, total power and network changes drops thanks to the reduced amount of hardware compared to legacy solutions. Global and per-port operational expenses are much lower than traditional solutions. Operations teams become more versatile and do not need to be allocated to specific networks, vendors, or technologies.
Disaggregated Networking Offers Cash Flow Improvements
The traditional way for building CSP networks is by buying monolithic chassis that are capitalized expenses over a period of 3-5 years. CSPs have been relying on their traditional vendors for building their networks with large upfront investments on equipment, services, operations and human skills, before the first dollar or euro is generated and regardless of any significant change in their future business environment (regulation, competition from other CSPs or OTTs, customer sales). History has shown this to be a risky approach. There is one alternative OpEx.
The OpEx model, widely used in other markets, like cloud services and enterprise software, reflects a change in procurement strategies based on risk-sharing, predictability, business agility and faster time-to-market through open ecosystems and designs. It’s also a proof point that every dollar spent must produce the intended result for the CSP – market share acquisition, new service launch, margin growth. If not, the CSP can stop spending or switch to another vendor in order to get the expected result. At DriveNets, we see more and more CSPs moving to the OpEx model, which may surprise the most conservative CFOs. However, we have also noticed that some operators are reluctant to move away from their traditional CapEx investment model. These operators can still enjoy a lower cash flow compared to their legacy solutions but not get the full benefit of disaggregated networking: low one-time hardware expense and low recurring software license fees.
Network Disaggregation Makes for Happy CFOs
Some leading CSP CFOs already see the advantages of the new network disaggregation model in their networks. Direct network costs are reduced thanks to less hardware, standardized form factors and a modular software-based model. Operations are therefore simplified, reducing OpEx spending and time-to-revenue. The payment method for disaggregated network costs is the CFO’s choice, but we see more and more organizations moving to the OpEx model like in many other industries. Network disaggregation can not only put a smile back on a network engineers’ face, but also make your CFO happy.
ACG Research White Paper
DriveNets Network Advantage - TCO Analysis