Most people (90%) can now get broadband over fibre connections so speeds are usually fast relative to those of yesteryear. But there can be a big difference between ISPs that is harder to assess than the comparison websites would have you believe.
There are three main choices: Virgin?s coax cable, ISPs who use BT?s fibre to the cabinet infrastructure, and small independents including satellite providers who are typically fast, high quality, and highest cost options.
Virgin?s headline speeds are typically faster than other provider?s but you need to delve deeper to find out the actual speed you would obtain in real life and at what cost. Virgin connections experience the greatest dip in download speeds at peak time (usually evenings); depending on your use you may not notice it always. Also, the upload speeds are often a third of the BT based ISPs? making cloud backup and photo uploading painfully slow.
BT based fibre speeds are offered in three tiers but the faster your tier the harsher the drop when the lines get busy. According to Ofcom of those customers on nominally 38 Mbps lines, 69% receive an average speed greater than 35 mbps but those on 76 Mbps nominal are worse as only 6% actually receive average speeds above 40 Mbps. This is often a result of the copper wire final stretch between the green cabinet in the street and the property?s master socket. So switching to a different ISP with higher advertised speeds will probably have no effect if you are far from the cabinet.
But different BT based ISPs are different in other respects: some will artificially throttle back the speeds depending on the price you pay or the amount of concurrent traffic passing through their equipment. For example, Sky and TalkTalk offer the cheapest 38 Mbps product for entry level customers whereas others use the middle tier 52 Mbps service. The upload speed difference is more pronounced: TalkTalk limit the top speed to 2 Mbps where many other are set to 10 Mbps.
One reason for the smaller ISPs being able to offer a faster service is that they better balance residential customers with business clients ? groups who have complimentary usage patterns. Also, they do not oversubscribe their backhaul lines as the greater price they charge enables better equipment to be leased.
However very small providers are not able to offer the content delivery networks whereby video (the highest demand media type) is served from content held locally where Google/Youtube, for example, will house a server in a local distribution hub to avoid saturating the backhaul infrastructure.
But in the current market most people buy several services from the same provider. Phone and broadband are often paired and this is usually a good idea to prevent one supplier blaming the other when a problem is experienced. The more additions though, the harder it is to judge the cost effectiveness of a package. This is related to customer service levels ? an area that is often not considered until a fault develops. For this aspect at least it is certainly well worth while taking advice from either an independent expert with wide experience, or the crowd. For the former I am too modest to make a suggestion but for the latter ThinkBroadband.com is the place to go for a consolidated view of thousands of real users. Note that seeking one or two opinions from friends is wasted time ? every company has a few disgruntled customers and a few who have never had cause to test out the features that are important to you. A PC Pro magazine study recently rated broadband providers for customer support, reliability, value, and speed and found that Zen and PlusNet were top of the heap, while Sky, EE, and TalkTalk took the bottom slots.
Finally, beware of charge creep. You should monitor the total package at the end of each contract, typically annually as a small increase in one facility is often unnoticed. Also, beware lockins: line rental for example is often good to pay annually in advance to secure the largest savings, but don?t buy this just before the end of your contract. Your negotiating position will be significantly weakened if they know you would be saying goodbye to over £100 of prepaid line rental were you to switch.
For many people there are two places where incoming email can be scanned and marked as probable spam, the repository that receives the message, and the mail client used to display it. If either gets it wrong you will either see messages that should be sent straight to the junk folder or have wanted messages hidden from your normal view due to being incorrectly flagged.
To deal with this don't just delete messages that should be in spam, train the software so that it makes a better decision next time. It is generally good practice to scan your spam folder periodically and when you get to hear of a message that was sent but not received so that you can improve the scanning algorythm for the future. If something is found don't just move the message to your inbox, tell the software that it got the designation wrong this time so that it will gradually learn what is and is not spam.
How you mark messages as [not] spam depends on which mail client (if any) you use and the particular mail repository. For example, if your mail client is Windows Live Mail you would right click a message to tell it about your preference. And if your repository is Gmail you would need to select the message in webmail and click the hexagonal spam icon.
This is feasible for most people but as my address is publically viewable on my website and has been active for many years, I get several thousand spam messages a week. Hence, I only check my spam folder when someone reports that I have not replied to their message.
This is really the sort of information better suited to my biannual newsletter, but as that is not due to roll off the presses until early December, here is where I place this pearl.
If you are sending or receiving email to/from anyone with an address that ends @BTxxxxxxx.com be aware that for several weeks (yes, weeks) they have been having intermittent problems judging by the very frequent automated bounce messages I receive.
Typically a message is not able to be delivered within a few minutes of it being sent. Usually (but not always) they are clearing the backlog within 4 days, but if you need a message to go faster than that you are best advised to switch to a provider that concentrates on email. In my opinion Google do a fantastic job.
The usual reasons for this are:
While here, those who receive broadband connections and possibly phone line rental from BT should check that they are receiving close to the speed you are paying for (which will probably be more than most other providers). Find out the estimated speed you should be enjoying at
Measure your current speed at
Or if you have a browser that will not work with that BT site then use
EDIT 4 Sep 16: I just sent a circular to 231 addresses. 30 bounced - all from BT domains.
EDIT 17 Sep 16: Hard to believe that the issue remains unresolved after so long. It seems that BT are very inflexible when handling spam abuse and recovery from it. I received the following advice from another ISP: "We have contacted BT internet to resolve this issue. However BT internet it not ready to share any information with us and they expect one of their users to contact them." So BT expect their users to know when they do not receive something then send the transaction log of the sender's ISP to them. Amazing.
It seems that ransomware is in the assurgent; not only are the reported (tip of iceberg of course) cases on the increase but the deft aplomb with which the customer service representatives of the crooks respond to requests for technical support indicates an industry that has come of age. The bad guys have a reputation to protect and they realise that by increasing their stature in the pit they inhabit they can gain a higher reward rate - folks who pay up.
It used to be the case that best practice was to never pay the bad guys who encrypt your family photos, work documents, or invaluable address book as you could never be sure whether you would actually receive the private key to unlock them and anyway paying only encourages them. The thinking is now that paying is better as the chances of them risking loss of reputation is that high, and the technical skills that good that the best outcome can be achieved by negotiating - albeit with a very weak hand.
More accurately, paying up is the second best option. Spending the same money on prevention is surely the better way to avoiding the stress that comes from one of these attacks. There are two avenues to prevent the problem in the first place, both are old chestnuts that offer protection far beyond the tedium that considering them may engender.
I am surprised at the number of computers I see running versions of Windows earlier than 10. I think for most owners it is fear of the unknown, belief in the (typically exadurated at best) scare stories, or indolence that results in this foolish state.
But the good, if only probably temporary, news is that you can still upgrade without having to spend the £85 or thereabouts that a Windows 10 license would normally cost. To do this you need to go to eBay (other auction sites are available) and buy a license key for Windows 7 - I just did this and paid £6.
Obtain the installation kit for Windows 7 from
https://www.microsoft.com/en-us/software-download/windows7 and install it.
Now visit Microsoft for a free upgrade to Windows 10 - currently available free for devices using assistive technologies. Go to
https://www.microsoft.com/en-us/accessibility/windows10upgrade for this from the computer needing the upgrade.
The only place left to visit is
http://www.1computercare.co.uk/feedback.php to sing the praises of this blog.
For years I have been advising people buying a new computer to shun all off-the-shelf outlets in favour of specialist online only retailers. See
I am comforted to read support for this position in this month's edition of PC Pro magazine issue 263. They sent 'blind buy' shoppers in to PC World and John Lewes branches and were left with the following conclusions
Their conclusion is that "computing remains an immensely complicated world" ... "Stick to the specialists". "In all you get much less for your money [with a branded product]".
Occasionally an internet connection drops periodically, this is normal and usually not noticed. But sometimes it drops frequently or regularly (or both). When this happens we can often diagnose it as being external factors that the internet service provider (ISP) will get sorted out for you.
Sometimes though the ISP reports that there is no loss of connection from their end. Now we have to look closer to home and by connecting to the master socket, thereby disconnecting the internal wiring asnd other appliances, it is often possible to confirm the location of the fault.
But if there is no fault seen after the above checks, and (usually) the problem is with wireless but not wired connections, then the possibility of Repetitive Electrical Impulse Noise (REIN) or Single High level Impulse Noise Event (SHINE) must be investigated.
There are very thorough descriptions of these events elsewhere (see rererences below) so here I shall limit myself to saying that interference can be caused by electrical items emitting a disruptive signal that can cause poor broadband performance. Not only dropped connections, it may be the ISP will automatically lower the speed as it prioritises stability over performance so the issue can manifest in different ways.
The solution is to make sure that your router is connected to the master socket and at least a metre away from all other electrical devices. This applies to the wires close by the router too as they can act as antenae that broadcast the interferance.
The test for this is to get a mediium wave band radio and tune it in to the frequency that causes the disruption: 612 KHz. Move the radio arond the router and you should hear the white noise much louder when close by the router and fade to nothing as you move a few inches away.
Examples of causes of REIN include:
More info from: http://www.kitz.co.uk/adsl/rein.htm
:: Next >>
|<< <||> >>|