« June 2009 | Main | August 2009 »
Is it time to change how broadband speed is advertised?
Ofcom's latest report into Britain's broadband speed, released today, reveals no surprises. Virgin Media's cable service takes the top spot, beating other ISPs, including BT, Tiscali and AOL, in some cases by a large margin. Ofcom has used the report to slam some ISPs because customers are receiving far less than the 'up to 8Mbit/sec' service they are paying for.
The figures only tell a small part of the story though, becuase broadband speeds are affected by many factors, some of which are under the control of ISPs, and some aren't. The two biggest reasons someone may not be receive the full speed their Internet connection is capable of are the quality of the line between them and their ISP, which includes the wiring not just in their house but also in the streets all the way to your telephone exchange, and contention with other users in the local area who are trying to access the Internet at the same time. A third is server load, since during busy times, a web server may have trouble handing out pages fast enough to everyone at once. Another is whether a broadband provider has placed a cap on download speed, which normally comes into effect once someone downloads a set amount of data over a given time period.
These issues aren't going away any time soon. The issue of line quality is the reason Virgin Media's cable service can deliver faster broadband than any other provider. ADSL is limited because it uses phone lines, many of which still use copper wiring, and in some rural areas of the country these were installed a long time ago. Until these are replaced, ISPs will never be able to provide the full advertised speed to all their customers all the time, and future speed reports will continue to look remarkably like this one.
Contention is an issue that can affect cable broadband as much as ADSL broadband, as it arises from an ISP not having enough network capacity in a given area to satisfy all its users. The issue can be resolved by an ISP digging into its pockets and upgrading its network and equipment. However, some ISPs cannot afford this, as retail prices for broadband have been driven down by cut-throat competition and price wars. Would people pay more for a service that can deliver advertised speeds more often than a cheaper provider? No, not everyone.
In truth, we're stuck with most of these problems for now, at least until fiber-optic cabling is installed to every home in the country, by which time we'll all have a few extra gray hairs, and the London Olympics will be a distant memory.
So if the problems cant be completely solved, a change in thinking may be better, and ISPs should sell and advertise packages that better reflect what they can actually provide rather than what they would like to. Why sell an 8mbit/sec service in an area where such a speed simply isn't possible? Why should someone living in that area who only gets half the advertised speed pay the same amount, for the same advertised service, as someone who does get the full speed in another part of the country?
The simple solution is to stop advertising speeds altogether. Perhaps pricing could be based on how much data you consume (like our electricity, gas and water supplies) rather than speed. Charge a small amount for light users, who may download email, surf the web and view a few videos. Power users who want to download considerable amount of data should be able to continue paying more for unlimited services. In all cases, the emphasis on speed in marketing literature should be changed. Is it beyond ISPs to simply tell you the estimated average speed for a given postcode area, and not be tied to this legally, rather than dangle the carrot of unattainable download speeds in front of your face?
By simply selling 1, 10, 50 or unlimited gigabytes of data per month, without explicitly saying whether that data will arrive at 'up to' 8Mbit/sec or 2Mbit/sec, ISPs will no longer be advertising something they cannot provide. Allow customers to top up their balance if they go over their limits, in a simple and affordable way, so they dont wake up one morning without access to email. Importantly, raise the speed caps (eliminating them probably isn't an option) and let the network infrastructure decide how fast data arrives. It does that anyway after all, which is why Ofcom became upset in the first place.
However this is unlikely to satisfy everyone, least of all me. I wouldn't dream of using any Internet service where my broadband is limited to a few gigabytes a month (a single game patch that downloads in the background would use my entire quota)
Why we may never see a Core i7 Apple machine
With the recent crop of new Apple desktop hardware - The Mac Pro, iMac and Mac Mini (all of which we reviewed in Personal Computer World) - we noted the use of Intel's older Core 2 Duo processor rather than the newer and more powerful Core i7 chip. The Mac Pro is an exception, but Apple's high end workstation doesn't use Core i7 either, as it's built around Intel's Xeon 5500 processor (which is similar, but not identical, to Core i7)
It's a certainty that Apple will upgrade its range to use Intel's newer processors at some point, but it may be the recently announced Core i5 and Core i3 that it chooses rather than Core i7.
If you haven't been following Intel's roadmaps, Core i5 will use a completely different socket to current Core i7 processors, using 1,156 pins rather than 1,366, meaning if you want to use a Core i5 processor, you need to fork out for yet another new motherboard. Intel's decision to bring another socket to the market has been met with confusion and some disappointment from those who have already invested in Core i7.
The good news is that Core i5 processors and components should be much less expensive than Core i7, but still offer high performance, partly thanks to an integrated memory controller that gives better memory performance than its predecessor.
This leaves Apple with a choice of processors for its next mainstream desktop line. We'll bet on them choosing Core i5 for the iMac, Core i3 for the Mini and sticking with Xeons in the Mac Pro, meaning LGA 1366 Core i7 processors end up never used in an Apple computer.
Of course, Apple hasn't even announced a new desktop line, and Core i5 processors aren't on the market yet, making this little more than speculation on our part. However, as long as Core i5 is relatively affordable and offers a reasonable performance boost, we'll call this an educated guess on what Apple's next move will be.
