Over the last couple of years, we’ve seen relentless growth in the consumption of Software as a Service (SaaS) and overlay offers in both the consumer and business markets. The movement started much earlier, but from a Telco perspective, the lines on the graphs are now moving at pace.

This blog will focus on the B2B market, however, the evolution we’re seeing in this market is already happening in the consumer market, where Over-the-Top (OTT) offers like Netflix and Prime Video require no complex engineering to provide high-quality content to our homes and devices. The design of IPTV networks previously required lots of complicated and costly design and equipment. Thanks to OTT services, this is no longer the case.

It’s been predicted for a long time that traditional services like Multiprotocol Label Switching (MPLS) and hosted voice would come under pressure from the SaaS/overlay movement.

So, the challenge for the Telco remains relevant, but more importantly, profitability. The challenge for the service provider is illustrated below.

On the left is the traditional approach which we designed to support the traffic profile at the time. This was broadly an 80:20 split of Intranet vs Internet, with fixed VPNs connecting users in offices and reasonably fixed locations connecting through a secure perimeter to applications in the private data centre. Added to this was a helping of Internet connectivity through a secured perimeter.

On the right is the new approach. It’s really not that new but a stark reality gave the pandemic acceleration of change. Devices and users are everywhere, applications reside predominantly on the Internet, and of course, security is required throughout.

The bottom line is that the intelligence and value of the network are being elevated out of the transport and into controllers that sit a layer higher and have a view across the entire network fabric.

The problem is that we have an architecture rich with features and capabilities to support the model on the left, which brings cost and operational complexity. The right side represents a panacea, but in reality, this end state will take time to get to. The customer cost expectation associated with the model on the right can’t sustainably be delivered using the model on the right.

So what do we do? At Cisco, we’ve thought a lot about this. If you strip back the costs associated with a network, it can be boiled down to a few key elements:

  • Up to 60 percent lower power consumption
  •  Up to 60 percent lower heat output (watts in usually equals power out)
  • Up to 90 percent footprint reduction

Lastly, we simplify and automate. Automating a complex environment is a somewhat vicious circle because the more complex the network the higher the automation cost becomes, thus negating the benefits. We’re seeing considerable advantages for many of our customers through automation, but as a standalone project, it’s not enough.

I’ve previously written about cleaning the house. I think it’s imperative to do this and set yourself straight for the next 10 years.

The network is more relevant today than it’s ever been, but it must be commensurate with the services it’s supporting. Service providers can reap the benefits of our innovation but they need to be prepared to be decisive and invest in the future. Maybe Cisco can help here too?

My recommendation is to be brave! The revenue and margins associated with the past are dissipating, so I encourage you to work with us to build an architecture and consumption model that allows you to deliver on the opportunities that undoubtedly exists for the service provider. We’re ready, are you?






Dominic Elliott

Chief Technology Officer

Global Service Provider - Northern Europe