How Fast Do We Need to Go?
A new study by eMarketer looking at the number of global broadband subscribers expected to happen by 2011 (567 million) got me thinking about how much bandwidth is enough to meet the needs of most consumers. Right now, 5Mbps seems to be the standard, although some South Korean ISPs are offering as much as 100Mbps.
Obviously, the need for an ultra big pipe is driven by demand for music, video games and video, particularly high-definition video. But it is becoming clear that the bigger the pipe, the higher the price for consumers. This reality has been seized as a marketing opportunity by the carriers and cablecos, which are slicing and dicing high-speed services into a variety of packages based on what consumers think they need. The flip side is ISPs have also been able to raise prices by emphasizing speed - something that resonates with consumers even if they don’t really need it.
What seems to be lost amid the focus on speed is whether high-speed networks can actually deliver at a time when video is eating up an increasing amount of bandwidth. This strikes me as a strange discussion when you realize it wasn’t that long ago that people were talking about a bandwidth glut and the existence of too much dark fiber. The amount of noise on the network has sparked a discussion about whether a new Internet needs to be created to complement the Web. This New Internet - which sounds like Canada’s CaNet4 and Internet2 in the U.S.- would see a network devoted to R&D activity, universities and other special projects.
In terms of the eMarketer study, the research firm expects the number of broadband subscribers will reach 567 million over the next four years from 285 million now. This will include 55 million fiber-to-the-home subscribers. eMarketer analyst Ben Macklin said a growing trend is broadband trade-up where consumers are migrating to “very high-speed connections. Plain old vanilla ADSL is quickly turning into the new ‘dial-up;”.
Update: According to a new report, N53% of US households now subscribe to a high-speed Internet service, and that high-speed now accounts for about 72% of all home subscriptions – compared to 60% last year.
June 5th, 2007 at 7:08 pm
Reminds me of a few Mark Cuban posts from a while back specifically on the development of multiple tiers of service(http://www.blogmaverick.com/2006/01/15/hey-baby-bells-amp-cable-we-need-multiple-tiers-of-service/) which I believe your kind of alluding to. Pretty much Cuban answers your question about why we need all this speed. It is due to content and furthermore the demand for this content. Everybody is dying to get the latest episode of 24 or Gray’s Anatomy in HD instead of standard.
June 5th, 2007 at 10:17 pm
Mark,
This is a lot of complicated questions wrapped up into one.
The “needed” bandwidth was little to do with the advertised speeds, that are occasionally reached in short bursts under good conditions. (Yes, the Korean 100Mbps service is much faster than the Canadian 5Mbps service, but neither will consistently deliver the maximum speed.) There are really two related questions:
1. What is the peak speed needed on a limited, but sustained basis? (i.e., to stream a MPEG4 HDTV show in real-time requires about 8Mbps for the duration.)
2. What overall usage/ demand do customers need? (This includes all their e-mail, surfing, VoIP, P2P, network backup, etc.)
Residential services do not support 24/7 usage of the network at the advertised speeds. In fact, they often don’t support 24/7 usage except at a small fraction of the advertised speed.
Bandwidth demand is growing in both dimensions at the same time. The set of applications dependent on higher speeds keep increasing (although, there are not significant apps beyond file downloads that push the current boundaries — even video streaming is usually designed to run at less than 1Mbps.) Also, the overall usage of the network is growing, particularly with applications (RSS, P2P, backup) that can operate in the background at relatively slow speeds but spread usage out over the day.
The other principal driver over the next few years (for telcos) will be the desire to deliver high-quality IPTV services. By design, this has nothing to do with the Internet, and actually places relatively low demand on the telco backbone network. It does, however, strain the capabilities of the local link, which must deliver multiple high-quality TV streams along with Internet access that can compete with the cablecos. (There are plenty of reasons to suspect that DSL technology will not be sufficient to meet this demand.)
There is no limit to how far this can grow because every improvement in connectivity will lead to new applications that can use it. (For example, expect the quality and resolution for videos to continue to increase as the customers’ ability to use it increases.)
BTW, there is still no shortage of dark fibre in the ground (but, it’s mostly in long haul, inter-city networks). The problem now, as it was in 2000, is getting the fibre lit and getting into locations (such as the home) where there is no fibre.