The end of the bandwidth race is here

CONSUMERS have long been abused by Internet service providers (ISPs) and their marketing departments. However it now seems that punters are wising up and looking for more than empty promises from the companies that connect them with the world wide web.

For a long time ISPs lived off the novelty of supplying an Internet connection that was not only faster than connecting through a modem but also avoided call charges, and that was enough to pull in the customers. As the Internet became a more immersive experience users started looking for more than just any broadband connection. But broadband had become a race, where cutting corners was common and catering for the lowest common denominator was the name of the game.

This country's major ISPs, most of which are run by large telecoms operators, would like to sell you an Internet connection solely based on its bandwidth. Virgin Internet advertises its fibre optic broadband while BT markets its premium ADSL offering as Infinity. Both of these firms are in fact selling a dream because the reality is that downlink bandwidth is not the answer for a high quality Internet experience.

Without doubt the downlink bandwidth of an Internet connection is a vital component in its overall performance, as after all a 2Mbps video stream cannot be viewed in real time over a 512Kbps Internet connection. But it is far from the only factor in determining end user experience. Though bandwidth was the choke point five years ago, today the web constriction comes in the shape of transfer quotas, traffic management and basic IP functionality - a view echoed by Carl Warner, proposition manager at Demon Internet, who says, "Other service metrics such as latency and packet loss are important."

As any network operator, engineer or researcher will tell you, the notion of a network devoid of resource limits is pure pie in the sky. But the sad reality is that networks are unable to cope with users' insatiable ability to consume data. Thanks to the unwillingness of most ISPs to deploy technologies such as multicast, anycast and IP version 6 (IPv6), among numerous others, consumers are left with second rate solutions that fail to address their fundamental problems.

Like a light bulb being flicked on, ISPs realised that marketing their products solely on bandwidth was the best way to hook customers. The only problem was, as bandwidth increased, growth in core network capacity didn't, and how could it? In the space of merely a decade not only has the last mile access link's bandwidth increased by orders of magnitude from 56Kbps, but the number of people connecting to the Internet has grown too. The ISPs had figured out the hard way that the decades old solution, over provisioning, simply doesn't scale technically or economically. The solution? Slap on prohibitive transfer quotas.

We rightly resent governments that introduce stealth taxes, but the tame mainstream media obsessed with bandwidth figures allowed ISPs to hide these figures in jargon filled policies. 50Mbps, sir? Yes of course, but only for two hours a day. However, Warner says that the industry isn't misleading customers by simply flogging broadband based on one metric. "That's the technology available so it would seem strange to artificially slow the service down." That might be so, but given the limitations on the use of the so-called state of the art technology, bigger isn't always better. However, there's change in the air.

Consumers are wising up and ISPs need to up their game, according to Warner. "Customers are definitely more educated in broadband services and connectivity." That education has resulted in deeper understanding with consideration being put on secondary functionality offered by ISPs. Cost isn't the primary driver for everyone, says Warner. "Customers are aware of cost but they are increasingly aware of performance as well."

Warner, unlike many of his industry peers, doesn't pull any punches when it comes to ISPs employing dodgy antics to subvert interaction between users and the service they purchased. "The difficulty for customers at the moment, is a lot of the small print with ISP's hiding the real service that is being provided. So customers are being shoehorned into products that may not be suitable for their needs."

As the utopian ideal of liberal transfer quotas seems unreachable, the emphasis should be on placing reasonable quotas, not ones that seem adept at merely serving as excuses for selling "add-ons" or other broadband trinkets. While Warner points out that Demon's transfer quotas start at 50GB per month, it is the upfront nature of clearly communicating what a company commits to giving its customers that can assure them of a good chance of enjoying the service without fear of additional charges.

With transfer quotas an evitable part of life, the question becomes, what happens if you are a heavy downloader? According to Warner, Demon's approach is relatively simple. "Very heavy users simply have their speed restricted during peak hours." And those worried that Demon will cut them off? Warner simply states, "We don't actually remove anyone unless they are doing something illegal."

While Warner's comments might not appease those who prefer to acquire their entertainment from Bittorrent and Usenet, that is not to be unexpected. As for the notion of bandwidth throttling, it might provide the best solution for those who don't want to worry about having to shell out extra just to remain connected to the Internet after one week's usage.

In effect the consumer broadband market is learning lessons from the business world. While the Delboys of the ISP world look for a smash and grab solution to making their money from consumers, businesses want a lot more than just a headline figure. Warner reels off a short checklist that makes one realise how almost infantile consumer broadband packages sound.

"Service metrics such as latency and packet loss are important with things like static IP addresses, assured rate connectivity and quality of customer service also counting," he said.

Even something basic such as a static IP address is a rarity, meaning users have to adapt to DHCP and perhaps traverse the black arts of network address translation (NAT). Even Demon's 'tenner a month' dialup service back in the 1980s came with a static IP address as standard, so why is it so hard to find something similar now? For ISPs more IP addresses don't cost a penny, since allocation bodies such as RIPE in Europe and ARIN in the US allocate ranges for free, providing you can justify a need for them.

It's true that without NAT the world would have run out of IP version 4 (IPv4) addresses a long time ago. However that doesn't explain why few ISPs and none of the major ones provide this basic functionality as standard. It leaves consumers and businesses paying for a basic service that costs absolutely nothing to the ISPs.

A maturing of ISP services is needed and much more than one with a higher set of numbers attached to meaningless marketing metrics. Broadband needs to be become more dependable with service level agreements (SLAs) placed on even consumer services. Why should consumers, who pay money, not be guaranteed service? After all, when you buy something at the shop you don't expect to hear that it might not work all the time, so why should this be the case with the provision of consumer broadband services?

ISPs for too long have managed to fleece consumers due to their ignorance. However it seems that as consumers start to realise that there is more to broadband than bandwidth, the ISPs will have to deliver much more than fancy ad campaigns.
While intuition says that competition should lead to better services, consumers are the ones that must drive ISPs towards providing truly useful broadband that has the ability to do so much more than just download small amounts of data very quickly. µ