White Paper: Is Internet cheaper than MPLS in the UK?

Long before SD-WAN came into the market, replacing MPLS access circuits with locally-procured Internet access circuits has been a method used to reduce costs for international WAN networks (see White Paper: Is MPLS dying?). As we discussed in our last white paper, A buyer’s guide to SD-WAN, there has been much hype around SD-WAN cost benefits and our paper analysed the reality of this as well as setting out other expectations. Since publication we have spoken to a number of organisations keen to understand whether these potential costs savings (as well as the benefits of SD-WAN) apply to the UK domestic market in the same way as they do for global WANs.

TNC has the benefit of reviewing thousands of price points every month across a wide spectrum of organisations and vendors whether through procurement activities, benchmarking, reviews of existing pricing or vendor offers. It is through this insight that TNC can see that, today, there is little or no difference between domestic WAN access pricing, whether it be MPLS or internet, on a like for like basis.

This fact often comes as a surprise to organisations that have been exposed to relentless SD-WAN marketing over the last few years purporting that internet access is cheaper than MPLS access. So, the question we will now examine is why is this the case?

Whilst TNC has seen lots of evidence of savings on international access circuit comparisons, the question that we are consistently asked is why this isn’t replicated in the UK? Having spoken to numerous vendors, and carried out our own primary research, TNC has concluded that this parity of pricing is influenced by several key factors:

• UK Telecoms Market – With telecoms deregulation in the 80s, the UK market has had decades of competition and remains very competitive driving both price and costs. Combining this with the density of population and in particular the predication to locate telecoms infrastructure in proximity to population means that there is a very high level of vendor and technology choice in most areas.
  
• Distance to POP/Exchange - The significant buildout of infrastructure in the UK means that access to private networks is relatively short in distance and it is distance and this dependency on PTT or equivalent that can add significant cost to a private MPLS network. If a vendor can reduce this dependence on 3rd party infrastructure, they reduce their costs. As such, for many providers, the distance from most customer sites within the UK to their nearest POP are relatively short.
  
• Network Running Costs - But surely, some commentators at this stage ask, even within the UK the cost of running a private network is more expensive than passing this traffic over the internet and thus never touching their “expensive” private network?
  

In reality, the costs of the core network are only one element in providing a service and can actually be quite a small constituent. The major cost constituents of setting up and running a WAN service are the same whether it be internet or MPLS – access circuits, engineering, project management, NOC, service and account management, sales and marketing, field engineering, legal etc.

Indeed, for most suppliers, their MPLS core networks are long established and largely already written off/paid for, save elements that are leased or recently upgraded.

In addition, assuming the costs of running a private core MPLS network are similar to the costs of connectivity to the public internet, for most vendors the MPLS running costs are internal whilst public internet connectivity (such as Linx) are external. Most vendors would prioritise internal cost over external costs.

In contrast to the position in the UK, TNC has seen significant evidence over many years that procuring internet services internationally is cheaper than buying MPLS. So, if this isn’t the case in the UK, why is it the case outside the UK?

In reality, there are many drivers for this but the underlying point is that the price difference is driven by material differences in the two services. An MPLS connection procured from a global network provider comes with a whole range of services and functionality in addition to the connectivity e.g. the functionality of MPLS itself (SLAs, guaranteed throughput, QoS etc.), as well as the service wrapper from the provider, including centralised billing, global contracts, global account and service management, language support etc.

By contrast, a locally-sourced internet connection typically has few SLAs, best efforts throughput, and little in the way of service wrapper, so billing, account management, service management etc. will all be local, often requiring local presence to contract, local language capability etc.

Some organisations will find these differences acceptable, and others will not, but they exist and they are a big part of the cost difference between these services.

Of course, these differences can also be seen in the domestic UK market. Most organisations with large UK networks procure them from UK market specialist carriers. However, they could instead choose to procure their UK network from a global provider that doesn’t specialise in the UK, and they would find that the costs would be much higher. Why? Because the circuits are being delivered with a service wrapper that costs money – for example contracting under Chinese law, helpdesk delivered in Japanese, account and service management in India etc.

There are other factors to also take into account:

In essence this point is more a reflection on the difficulties for global telecoms companies of providing MPLS circuits in some countries and the supporting local tails in comparison to the UK. The telco needs to establish local relationships and, whilst this should be bread and butter for them, there will be some countries where this is difficult, and/or they have limited economies of scale due to low demand for services (think unusual territories). Compare this to the ability of an organisation who can buy locally, is familiar with the local services, speaks the local language etc etc. So, whilst buying and providing services in France or the US might be familiar and easy, the

ability to buy services in Papua New Guinea or Laos might be a very different experience. In these circumstances it is not surprising that buying a local internet service may be far cheaper than relying on your global MPLS provider to provide and buy (centrally) an equivalent MPLS service.

Some countries are less densely facilitated by telecoms infrastructure than the UK which results in longer tail circuits and a less competitive market - indeed some remain governmental monopolies. The cost from the local provider to the international MPLS provider may be equivalent to their local internet price but the international MPLS provider then has substantially more costs to add, not least they need to backhaul the local access on to their network. Some POPs are not even in country due to the low level of demand they have for services to these countries – again take the example of Laos the nearest POP might be Bangkok or even Hong Kong whilst for Papua New Guinea its likely to be Sydney. Thus your global MPLS provider is paying a local provider for an international tail as well as the access circuit as opposed to your local internet provision which is locally connected to the global internet backbone.

As such global or regional procurement of services adds cost to a multinational provision in comparison to local internet procurement (not just added costs and margin stacking but this may be exacerbated by a poor understanding of the local market, poor buying and costs of translation/local language support).

Within the UK, TNC currently sees no evidence to suggest that Internet-based WANs are any cheaper than MPLS-based WANs. There are multiple reasons for this but most importantly it represents the competitive nature of the UK market, and the similar comparative costs of running either type of network.

Other than matters relating to The Network Collective, this research is based on current public information that we consider reliable. Opinions expressed may change without notice and may differ from views set out in other documents created by The Network Collective. The above information is provided for informational purposes only and without any obligation, whether contractual or otherwise. No warranty or representation is made as to the correctness, completeness and accuracy of the information given or the assessments made.

This research does not constitute a personal recommendation or take into account the particular investment objectives, financial situations, or needs of individual clients. Clients should consider whether any advice or recommendation in this research is suitable for their particular circumstances and, if appropriate, seek professional advice.

No part of this material may be (i) copied, photocopied or duplicated in any form by any means or (ii) redistributed without the prior written consent of The Network Collective Limited © 2019