The cost of Internet access – and the importance of IP transit

Data recently released by the International Telecommunication Union states that achieving universal access to broadband connectivity before the decade is out will cost about $428 billion. It’s a staggering figure that considers infrastructural needs, basic digital skills, local content and regulatory frameworks.

Africa’s broadband ecosystem certainly stands out in terms of need and investment potential. The continent is home to 17.2% of the world’s population and yet less than half of its inhabitants use the Internet.

This presents great opportunities for collaboration towards business growth and economic development. However, it’s important not to undercut on quality – particularly as the coronavirus pandemic has shifted so much of our lives online, including access to essential services.

Addressing the digital needs of underserved communities is vital work that has the potential to transform people’s lives.

I’m fully behind the idea of digital connectivity as a human right – but that doesn’t mean any low bit rate connection will do. Everyone has a right to reliable, high-quality and affordable connectivity – and that needs to be our baseline.

Interestingly, when we look at all the components involved in broadband service delivery, an unexpected (for some) picture emerges. The costliest component is not actually the one that impacts the quality of the service the most. Take a look.

The IP transit component has the biggest impact on the end-user; it allows users to send and receive packets of data to every other connected device in the world reliably and at the highest quality possible (taking redundancy, latency and routing into account).

Yet, at around 3%, IP transit is a fraction of the overall cost for an ISP to provide an Internet service to an end customer (in South Africa). However, the cost of the last mile gobbles up almost 60% of the total revenue! This will naturally vary depending on the cost of IP transit in the market, regulatory fees, the number of last mile providers, and so on, but it does provide some important food for thought.

I’m fully behind the idea of digital connectivity as a human right – but that doesn’t mean any low bit rate connection will do.

All too often, ISPs focus solely on price, and don’t consider the question of quality as much as they should. ISPs shouldn’t risk buying the cheapest IP transit on the market – sinking the ship for a ha’penny worth of tar and all that. Instead, they need to see it as an area that needs maximising for the best, most cost-effective results.

Prioritise network security: The Mutually Agreed Norms for Routing Security (MANRS) operator community helps ensure groundwork is in place to acceptable levels of security, stability and performance within the Internet community at large. So, ideally you want an IP transit provider that implements the MANRS routing manifesto actions (this can be checked here). Another critical security element to consider is RPKI ROV; I would strongly suggest making sure your upstream ISP drops invalids (which you can do here).

Double-check the local routes: It’s important to know if your IP transit provider is well peered with all of the major local networks (both content and eyeball networks), preferably over PNI. That way, they can keep traffic local if an Internet exchange has a wobble, or there is a major failure network. The most reliable way to verify this is through a looking glass; if they don’t have a publicly accessible looking glass listed on PeeringDB, that’d be a big red flag for me.

Ask for guarantees: Like any good partnership, make sure your IP transit provider operates transparently and is not afraid to stick their neck out for you. If there’s an attack, can they work with you to mitigate its impact? Do they have enough highly skilled engineers to work with you through major failures? What are their latency and packet delivery guarantee levels? A financially punitive SLA provides confidence that the provider is going to go above and beyond to get the job done.

Renew relationships:In a time of ‘great deals’ and ‘great prices’ don’t underestimate the importance of great long-term relationships. Even if the price-tag is slightly higher, I think it’s wise to consider the value of a proven quality service that comes with 24/7 support and in-country engineering expertise that you can build a strong relationship with over time.

In pursuit of ubiquitous connectivity across Africa, let’s not drop the ball on quality. While cost is always a necessary consideration, knee-jerk decisions based on price alone are rarely successful. When it comes to IP transit, they make no financial sense at all.

Let’s all do our bit to keep the African traffic in Africa as opposed to looping it out of the continent and back again for no good reason at all.