高速增长的数据流量是否会压垮无线网络？（zz）

Saturated mobile networks
满负荷的移动网络

Breaking up
崩溃

Will the rapid growth in data traffic overwhelm wireless networks?
高速增长的数据流量是否会压垮无线网络？

Feb 11th 2010 | From The Economist print edition



BEFORE the end of next year the internet will collapse, predicted Bob Metcalfe, a noted network engineer now turned venture capitalist, in a magazine column in 1995. As traffic grew exponentially, outages proliferated and telecoms firms seemed overwhelmed. But he was swiftly proved wrong, and literally ate his words by mixing a copy of his column with water in a blender and swallowing it. Yet as the industry gathers in Barcelona for the Mobile World Congress, its biggest annual shindig, it feels like 1995 all over again. Wireless data traffic is exploding. In some cities capacity is already running short. Pessimists are once again giving warning that operators must rein in usage or face collapse.

1995年，鲍勃·梅特卡夫，一位已经投身于风投的著名网络工程师在一个杂志专栏中预测说因特网将在第二年年底前崩溃。数据流量成倍地增长，通信故障频 发,这都使电信公司有点不知所措了。但是他的想法很快被证明是错误的，他确实承认了自己的错误，将自己的专栏副本和水一起放进搅拌机吞了下去。但是当移动 通讯界的代表相聚在巴塞罗那的移动世界大会，这个业内最大的年度聚会时，仿佛又回到了1995年。无线通信流量正在爆炸式地增长，一些城市的承载量已到极 限。悲观者再次提出警告，运营者必须对使用量严加控制，否则产业将面临崩溃。

In the mid-1990s cheaper personal computers, faster modems and the birth of the web drove demand. Today it is the success of smart-phones and devices known as “dongles”, which connect laptops to mobile networks. At the end of 2008 there were 189m mobile-broadband connections, generating on average 175 megabytes of traffic per month, according to Bernstein Research. A year later the respective figures were 312m and 273 megabytes. Data traffic thus grew by a whopping 158%.

90年代中期带动需求的是廉价的个人电脑，快速的调制解调器和新生的网络。今天的则是智能手机和连接手提电脑与移动手机网络的移动外设装置。根据伯恩斯坦 研究公司的数据，2008年底，1.89亿移动宽带用户平均每月产生175兆字节流量，而一年后数据分别为3.12亿和273兆字节。这样，数据流量取得 了158%的巨大增长。

Just as in the early days of the internet, there is heated debate over whether such growth will continue. Many of the same people are pitching in. The mobile internet will “ramp faster than the desktop internet and will be bigger than most think”, argued a recent report overseen by Mary Meeker of Morgan Stanley, who was one of the most bullish analysts of the dotcom bubble. Cisco, which has a vested interest as the world’s biggest maker of network equipment, expects mobile-data traffic to increase 39-fold over the next five years. Andrew Odlyzko, a professor at the University of Minnesota who correctly predicted a slowing of the growth in fixed-internet traffic, is again more cautious, expecting growth to slow over the next couple of years.

就像当年的因特网一样，这样的增长能否持续引起了热烈的争论。参与争论的有不少熟悉的面孔。摩根斯坦利的玛丽·米克，这位在网络泡沫时期是最狂热分析师之 一，在其监督的一份报告中指出移动网络将以超过因特网的速度加速发展，规模将超出任何人的想象。思科，这个利益相关者作为全球最大的网络设备提供商预测移 动数据流量将在今后5年增长39倍。而曾正确预测出固定网络数据量增长将放缓的明尼苏达大学教授安德鲁·奥德林克这一次则更为谨慎，他预计增长将在今后几 年放缓。

Meanwhile operators are scrambling to catch up. The “huge irony”, says Ben Wood of CCS Insight, a market-research firm, is that only five years ago the industry was scrabbling for customers for the new networks in which it had invested billions. Operators cut capital expenditure and offered unlimited, flat-rate plans, only to be caught out when demand finally took off.

与此同时运营商们在追赶网络成长脚步上却略显疲态。仅仅5年前这项已经投入数十亿的产业还在为新网络摸索消费者，市场调查公司CCS Insight的本伍德认为这是“巨大的讽刺”。运营商削减了资本开支并提供了无限制的包月方案，却在需求最终起飞时没有做好准备。

AT&T, America’s second-biggest mobile operator, is just the most visible example. To secure exclusive rights to the iPhone, it acceded to Apple’s demand that the device come with a simple, flat-rate plan. But instead of selling only a few million as it expected, AT&T now has an estimated 12m-14m iPhones on its network. Its data traffic has grown by 5,000% in the past three years. Small wonder that it has had a hard time coping, particularly in such technophile cities as New York and San Francisco. Things have improved recently, but the network may clog up again if the iPad, Apple’s new tablet computer, for which AT&T is also providing mobile access, proves a similar success.

美国第二大的移动运营商AT&T就是最明显的例子。为了取得iPhone的独家运营权，其同意了苹果公司关于按时计价的简单方案。但是 AT&T现在的网络中拥有1200到1400万的iPhone用户，这大大超出了其仅仅销售几百万部的预期。其数据流量在过去三年增长了50倍。 难怪AT&T曾一度疲于应付，特别是在移动技术爱好者集中的城市像纽约和旧金山。情况最近有所好转，但是随着AT&T提供网络接入服务的 苹果公司新的平板电脑iPad的诞生，苹果产品的类似成功可能会使网络再次陷入拥堵。

The looming capacity crunch of the 1990s was averted not just by slower demand growth, but also dramatically increased supply, in the form of big investments in fibre-optic cable and technical tricks that squeezed more data through the pipes. Mobile operators are duly beefing up their networks and preparing to roll out LTE, the next generation of wireless technology. Some, such as Vodafone in Europe, continued to invest heavily during the recession because they regard network quality as an important selling point. AT&T, meanwhile, plans to increase its wireless investment by $2 billion this year. And industry observers such as Ameet Shah of PRTM, a consultancy, expect consolidation to spread the burden, either through sharing of networks or outright mergers, as Orange and T-Mobile plan in Britain.

90年代潜在的带宽危机得到避免不仅仅是因为放缓的增长需求，还有增长巨大的供应能力，投入到光导纤维中的巨资和技术改进从电缆中榨取到了更多的数据带 宽。移动运营商正适时地加强他们的网络技术并准备推出下一代无线通信技术LTE。一些公司，诸如欧洲的Vodafone公司，在萧条期也持续投入巨资因为 其将网络质量视为重要的销售点。同时，AT&T也计划将其今年的无线技术部分的投资增加20亿美元。业界的观察家，像PRTM的咨询师阿米特·沙 阿预测加强合作是填补费用负担的手段，要么共享网络要么完全的合并，就像Orange和T-Mobile在英国的计划那样。

But wireless technology is not improving as fast as the kit behind fixed-line internet services did in the 1990s. The number of mobile connections in America has ballooned from 2.7m in 1989 to 277m in mid-2009. But operators have coped with this not by using spectrum more efficiently or acquiring more of it, for the most part, but by putting up more radio towers: their number increased from about 3,600 to nearly 246,000. Experts predict that to meet the growing demand, operators in developed countries may have to treble the number of base stations.

但是无线技术的提升并不像90年代固定线路网络服务设备那样发展迅速。美国移动用户从1989的270万膨胀到2009年中的2.77亿。但是运营商应对 的方法并不是更有效地使用频谱或是提高其利用率，而是建造更多的通信塔：它们的数量从大约3600上升到近246000。专家预计运营商为了满足需求，其 在发达国家可能需要建设的基站数量是前一数字的三倍。

Moreover, wireless networks are different from fixed ones. In most cases, the extent to which one subscriber uses a fixed connection has little impact on other customers, since each has a separate link to the internet. But the data-transfer capacity within a cell is shared between several handsets. If left unchecked, a small group of users can gobble up most of the bandwidth, as on AT&T’s network, where the top 3% of users consume 40% of it.

此外，无线网络和固定网络也有所不同。在很多情况下，某个用户使用固定通信的程度对其他的消费者影响很小，因为每个人都有一条单独的网络连接。但是手机的 数据传输能力是在几个通信接收站内分享的。如果不受控制，一小部分使用者将消耗掉大多数带宽，就像AT&T的网络，3%的用户消费了40%的带宽 总量。

No wonder, then, that mobile firms do not believe in “network neutrality”, a much-cherished principle of the fixed internet, which holds that operators should not play favourites with certain kinds of traffic. AT&T reserves the right to cut off heavy users of file-sharing, which is thought to account for almost two-thirds of data traffic on some networks. Even so, thanks to prolific users on flat tariffs, expanding capacity will not necessarily bring in higher revenues.

移动公司不坚守“网络中立性”并不奇怪，这项原则在固网中备受珍视，主张运营商不能对某些特定客户流量有所偏好。AT&T保留了切断大份额用户关 于文件共享的权利，这部分被认为占到了某些网络数据流量的几乎3分之2。即使如此，归功于大客户的固定包月方案，扩大的带宽容量并没有必然地带来更高的收 入。

The politics of wireless networks are also different. The cheapest way to increase capacity is to add more spectrum or to move a network to a lower frequency, which allows radio waves to penetrate walls more easily. So operators tend to lobby governments for more and better spectrum before investing in expensive kit.

无线网络中的中的政治手法也不一样。提高带宽容量最便宜的方法是增加更多频谱或是某个网络降低到更低的频率以使无线电波更容易穿透墙壁。因此运营商倾向于在昂贵的设备商投资前先去政府游说以期获得更好的频谱资源。

Another disincentive to investment is the threat data traffic poses to the industry’s cash cows—voice calls and text messages—which still generate 85% of revenues. Increased data capacity, after all, makes it easier to use alternatives such as instant messaging and Skype, which could become as pervasive on smart-phones as they are on personal computers.

另一个针对投资的抑制因素是数据通信对产业的现金牛造成的威胁——语音通话和短信服务仍然占到85%的收入。提升的数据容量终究使其他工具例如即时消息和Skype变得更容易实现，这些工具将在智能手机上变得和在个人电脑上一样普遍。

Operators will try to manage traffic in all sorts of ways. One tactic is to offload it to the fixed internet: the iPhone, for instance, switches to a Wi-Fi network whenever possible. Another is to try to get households to install what are known as “femtocells”—wireless base stations for the home. Many European operators already reduce connection speeds for the greediest users. Telefónica, one of the world’s biggest mobile operators, recently said that it is considering charging Google and other big internet firms for access to its network.

运营商将会以各种方式掌控流量。一种手段是将一部分分流到固网中去：以iPhone为例，它可以在任何可能的时候转接到Wi-Fi网络。另一种是尝试让家 庭用户安装所谓的“接入点技术”——无线网络基站的家庭版。很多欧洲的运营商已经降低了流量需求过大用户的连接速度。Telefónica，世界最大的移 动运营商之一，最近宣布将会对Google和其他互联网大公司接入其网络收费。

Yet all this may not be enough, argues Craig Moffett of Bernstein. Sooner or later, he says, operators will have to introduce more stringent usage caps, demand a premium for better service or charge by usage. That will be unpopular, judging by the public outcry when AT&T recently said it was looking at ways to curtail data traffic. “Users would like fixed and mobile internet to be the same, but the huge capacity gap between fibre and radio links is unlikely to ever allow this,” says Mr Odlyzko. Mr Metcalfe, meanwhile, may have to eat another of his columns. “It is my sad duty to inform you”, he wrote in 1993, “that the coming resounding flop in wireless mobile computing will be, alas, permanent.”
伯恩斯坦的克雷格·莫菲特认为虽然上述手段可能仍显不够，但运营商迟早会引入更严格的使用上限标准，对更好的服务和消费收费。这会是不受欢迎的，以最近 AT&T宣传将寻求渠道削减数据流量为例，公众表现出不满。“用户希望固定网络和移动网络一样，但是光纤和电波间巨大的容量差距使这愿望不太现 实，”奥德林克说道。而梅特卡夫可能要去吃掉他在1993年写下的另一份专栏了，“我非常伤心地告诉你，接下来无线移动计算机的巨大失败将会是，唉，永久 的。”