Who Pays?

In fact, a more accurate title to this blog post might be, “Who Pays and Who Suffers?” or possibly, “Who Pays but Still Suffers?” I guess this could be applied to many things (including the wonderful tax system here in the UK!), but what I am referring to here is the internet and more specifically how access to it will be provided in the future.

It’s common for us to use the graph to the right when talking with analysts, network equipment providers and carriers; it shows the growing gap between traffic volume and subscriber revenue. The recurring question is how the gap could be closed with such strongly diverging and yet fundamentally linked parameters.

There are numerous reports including the Cisco Visual networking study that shows bandwidth roughly doubling every year. Video is often named as the killer application, or even the network killer.

Fundamentally one of two things needs to happen: revenues need to grow to kick up the orange curve, or else traffic needs to be limited to reduce its rate of growth. Tiered services are proposed to lift Average Revenue per User (ARPU). But let’s be realistic, any growth in ARPU is going to be much smaller than the 2X increase in capacity required each year – so tiered services can only ever be a small part of the answer.

And, equally, shared revenue models from advertising and priority for specific services can only be part of the answer too. Why? Number 1: Do you think the content providers will give up these dollars? It’s unlikely at this stage. Number 2: This would play straight into the net neutrality debate if one service suddenly got preference over another.

OK, next plan then: let’s rate limit traffic to bring the blue curve closer down to the orange one. This is a good short-term defensive play and one many operators – especially in the mobile space – are trying. Sales of Deep Packet Inspection (DPI) equipment deployed on the Gi interface in a mobile network are growing quite nicely, thank you very much. Policy enforcement will always be critical, but there’s a point where it will stifle innovation and use of the internet. So again, this can only be part of the answer.

Try again…traffic in terms of Gbps or whatever is appropriate at your point in the network can also be considered in terms of dollars/Gbps. Moore’s Law! Great, we get a 2X performance increase roughly every couple of years. This is much more in-line with needs.

Enter 40G ATCA, 40GbE & 100GbE Ethernet and a whole raft of new processor technologies including switch, compute and packet processors. This is one of the fundamental ways to reduce cost per bit and close the gap. And given ATCA is an open ecosystem with multiple vendors, unlike our proprietary blade compatriots, you can be sure of staying on the leading edge.

Finally, I am seeing a fundamental shift in the way customers are approaching the decision process. Suddenly cost is Number 1, and having the “perfect” architecture is Number 2. The accountants and commercial people now have a bigger say in the selection, whereas before I sensed the architects held the balance of power.

What this means for vendors is that we will see more inventive applications of ATCA and technology to reduce cost. For instance, we have just worked with one customer to propose a new approach…with the application of new technology and different configurations it is possible to reduce the cost by almost 60%. Now this gets really interesting…

In closing, there is no single magic bullet that fixes the divergence in traffic growth and revenue growth. A blended approach is the only path, and technology innovation will play the most important part to reduce cost per bit. Getting the right technology, a cost-effective architecture and the right partner will pay the biggest dividends. Give us a call!

Twitter icon