Cisco’s Visual Networking Index (VNI) is an annual forecast of Internet traffic trends that has become a staple of the content delivery industry. It’s hard to go to a conference or attend a webinar without seeing Cisco’s data quoted, along with the projected hockey stick of traffic volume, with the majority being video. Two months ago, Cisco released their latest data and viewed from the 30,000 foot level, the narrative has not changed.
Consumer Internet traffic continues to grow dramatically, at a 32% CAGR from 2011 through 2016. In fact the forecast itself has grown; in 2011 Cisco projected total 2012 consumer Internet traffic at 24,476 PB/month, but this year’s report shows 2012 at 30,034 PB/month, a 22% upward adjustment in just one year. Not surprisingly, Internet video represents the largest single portion of this for the entire forecast period.
For the most part all this is still true, but the full story is not quite so simple. As anyone who follows the CDN space closely knows, service based CDNs don’t make a lot of money delivering video. Margins are small, most CDNs aren’t profitable on their video traffic and all CDN vendors are working hard to diversify their revenue away from video only products.
For video being delivered on-net, or inside the MSO, ISP or carrier network, these kinds of numbers from Cisco are nectar. There is seemingly no end to the growth in video traffic flooding operator networks, so there’s a constant need for products that compress, cache, offload or optimize video traffic, inside the last mile. And since operators are mostly deploying these solutions for cost savings and QoE, with monetization models to follow, vendors that provide these type of platforms inside the network are seeing more growth than service based CDNs.
While it’s true that video is the largest single source of traffic, it’s far from dominant. At one point the report states that “the sum of all forms of video (TV, video on demand [VoD], Internet, and P2P) will be approximately 86% of global consumer traffic by 2016.” However, this number includes a very large portion of traffic such as peer-to-peer for which the underlying content is assumed to be video but the network traffic is a non-streaming form such as a file transfer, which does not lend itself to video compression, optimization or video-only caching.
The portion of consumer Internet traffic that is truly streaming video in 2012 is 57%–still significant but far less than the 86% headline–followed by file sharing at 24%, and web, email and data at 18%. What’s more is that the video portion is forecasted to decline slightly to 54% by 2016 (although it continues to grow in absolute terms). If you look at the market today, video-only tools address at most the 57% (declining to 54%) of the total traffic. Meanwhile P2P traffic tools are not thought of as “video” tools at all, even though most of the time the content they help move is video. It’s a fragmented picture and “80/20” rules do not apply.
We also constantly hear people talk about the growth of mobile data traffic and how mobile is overtaking fixed line traffic. However, according to the VNI numbers, mobile is still a very small fraction of total bits transmitted, comprising 3% today and growing only to 10% by 2016. So while mobile is important, it’s not as big of an opportunity as some make it out to be. In addition, most mobile video consumption takes place over WiFi, not over cellular. Video over 3G and 4G is not taking off as the pricing for these services are far too expensive and carriers have yet to come to the reality that consumers will not pay what they want. Mobile video consumption, especially for long-form video content, is at a stand still not because of technology, but because of the current pricing model and cap limits imposed by carriers.
So what is the implication of all this? From a traffic management perspective, operators need to deal with video but they cannot afford to have a video-only or mobile-only traffic strategy. Operators who want to provide a high quality of experience for their users and maximize offload must deploy solutions that look at the entire scope of content traffic. We tend to gravitate to 80/20 rules because they make our lives simpler, but Cisco’s latest VNI provides a case where the 80/20 rule does not apply. Effective solutions for delivering content need to address the reality of what’s really taking place in the market today and not a simplified story.
I haven’t been blogging too much over the past few weeks as I’ve been recovering from arm surgery which hindered my typing and blogging a lot more than I thought it would. Apparently I’m not as superhuman as I thought. The good news is that I’m well on my way to a full recovery now and can get back to typing for hours on end.
