Barely a week passes without new analyst figures announcing the continued fierce growth in bandwidth and the concern that our networks may soon run out of capacity. This week it was the turn of IDC to release its findings on the state of the networking industry. IDC’s figures confirmed previous studies that show the dramatic growth in broadband connectivity. It expects to see Internet users to reach 2.7 billion by 2015. This represents over 40% of the world’s population and nearly a billion more users than in 2010.
Growth on this scale is phenomenal. What’s more, IDC expects most of this growth to be in mobile broadband, especially as tablets become more affordable (check out an earlier post on this: Mobile Data and the Ticking Tablets). In many respects, this echoes Cisco’s Virtual Networking Index study that predicts a 26-fold increase in mobile data between 2010 and 2015. What’s clear is that mobile usage will overtake PC usage and that mobile data is only going in one direction.
However, although we often hear about these dramatic increases in bandwidth demand in this post-PC era, we rarely hear about how the industry is responding. This thought was made quite clear by a reader in an earlier blog post: see Will Video Become a Network Killer? In many respects this comment is true. We often highlight the issue but not the solution. Yet looking around the industry, it’s clear to see why – there aren’t many. Most vendors seem intent on selling products to service providers and enterprises that are too expensive and too short sighted.
The drive towards 100G networks highlights this situation quite clearly. Earlier in the year I wrote about the optical reboot and how we are at the first stage of seeing our global networks being rebuilt on a foundation of 100G, OTN and ROADM technology. While this work appears to be continuing in the core, the metro has been completely overlooked and this is something that cannot happen. The metro is critical in the transportation of data for residential broadband, mobile backhaul, business Ethernet and data centre connectivity. These are some of the most vital applications in the network and cannot be neglected.
The main issue here is that current 100G long-haul solutions do not scale down to the metro. For distances of up to 500km, coherent 100G cards are simply too expensive and, hence we only have a partial optical reboot: a two-tier network where bandwidth is not evenly distributed. If we want to see our networks develop and accommodate the massive growth in broadband demand, we need to see an even spread of 100G throughout the network. The only way to effectively achieve this is to adapt 100G technology to fit the specific needs of service providers and enterprises. In today’s networks, one size does not fit all.
To do this, we need to see non-coherent 100G solutions with direct detection in the metro. Indeed, some of the biggest names in the enterprise community are already demanding them. Walk into any data centre serving the metro today and they will tell you that they want 100G that’s cheaper, smaller and less power hungry, yet more spectrally efficient than today's 10x10G solutions. These are the key ingredients for a 100G metro solution. So far, there is only one solution on the market that ticks all of these boxes. Other vendors seem intent on trying to scale down coherent solutions or using other techniques that simply don’t match users’ needs. It’s simply a case of a round peg in a square hole.
What’s clear is that if we are to successfully address this continued growth in broadband demand, whether it be fixed line or mobile, the vendor community needs to do a much better job of working closer with service providers and enterprises to provide solutions they need rather than solutions with a bigger margin. What inspires me about working in the tech industry is finding companies that actively want to do this; companies that want to innovate and drive the network forward. Only by collaboration can we truly develop a networked world. A world that can respond to the enormous challenges we face in a unified and connected manner.
What are you seeing in the network? Do you believe that we need both coherent and non-coherent 100G solutions to answer the diverse networking demands? Are we closer to seeing networks scale to meet the growing bandwidth demand? Let me know what you think on this one.