Internet data usage: Up, up, and away, and 'going mobile,' as The Who might say
In a recent write-up, I discussed how traditional POTS telephony was rapidly becoming irrelevant, due in part to a wired transition to VoIP telephony and in part to an exclusive reliance on wireless cellular connections. And I noted that in the latter case, legacy cellular voice connections are even being supplanted by audio-only and audio-plus-video cellular data services such as Skype and Apple's FaceTime. All of those data packets, particularly those of a video-inclusive nature, put an increasingly notable burden on IP networks. And as such, I always look forward to Cisco's annual Visual Networking Index Report. No matter how much I might already be familiar with the underlying trends, I'm still mind-blown when I see the numbers.
EB equals exabytes, folks. An exabyte is a 1 followed by 18 zeros. It's alternatively stated as one quintillion bytes, or a billion gigabytes, or a million terabytes. Here's your visual aid:
And keep in mind that the above graph's units are per month. Multiplying by 12, this means that at some point between 2014 and 2015, the global IP traffic metric will cross through the 1 zettabyte (1x1021) per year threshold.
I can brainstorm at least three primary factors which are driving this impressively increasing usage trend:
- Human population growth over time
- A growing number of devices used per person (and specifically, the number of per-person devices that are simultaneously sending and/or receiving data), thereby rationalizing the urgency of the in-progress IPV4-to-IPV6 transition, and
- An increasingly large per-data-transfer payload size
To point #3, video is the disproportionately dominant factor driving per-device data usage growth. As I've observed on numerous occasions, most recently in mid-June, the era of physical media-housed music has largely drawn to a close, with the video equivalent in the process of following it into extinction. Netflix was an early advocate of the streaming video successor, and although the company may have been a bit over-enthusiastic with its disc-to-download advocacy in the short term, its long-term instincts were sound.
Equally intriguing to me in Cisco's latest report is the dramatic fixed-to-portable equipment transition suggested by the data.
As Ars Technica's coverage notes, "In 2011, Wi-Fi accounted for 40% of all traffic, while cellular data accounted for 2%, with the remainder coming from traditional fixed traffic. ... By 2016, Wi-Fi will account for 51% of all traffic and cellular will quintuple, moving up to 10%." Keep in mind that Cisco's report tallies only the initial connection from a client to the network; if a laptop links up to a router over wireless 802.11, but the router connects to the Internet over wired DSL or cable, the packets aren't double-counted; it's a Wi-Fi "vote."
Admittedly, some fixed clients otherwise capable of being wired-connected via CAT5e (or MoCA or powerline, for that matter), such as desktop computers, are instead tethered over Wi-Fi solely for convenience reasons. But I suspect it's a safe bet that the bulk of those wireless-connected clients are portable devices capable of running solely off batteries. And I strongly suspect it's a safe bet that anything cellular data-connected to the Internet is a portable device (laptop, netbook, tablet, cell phone, Ebook reader, etc).
In the near future, I'll publish a follow-on post discussing the need for sustained and substantial increases in broadband bandwidth, along with more sanity with respect to bandwidth caps and overage charges, to make Cisco's prognostications a reality. But for now, I welcome your thoughts on the topics I've raised here.