Posts by Corin 45 posts • joined 19 Dec 2007
I use uMatrix, set to disable javascript everywhere but it allows me to express which domains can load javascript on which sites (so I can allow first-party JS on a particular domain)
I think that strikes a sensible balance; by default everything is blocked apart from what I specifically allow.
That could well explain the behaviour; if (for example) Talktalk have private peering to AWS and CloudFront then that link being at 100% might mean that Prime Video, Twitch and AWS-Hosted sites would be slow, while services hosted via other network links into TalkTalk would still be seemingly fine.
Where's this 5-8 year figure keep coming from? Have a look at the following article:
https://electrek.co/2018/04/14/tesla-battery-degradation-data/
It seems that total and utter battery failure is very rare, and the more common degradation due to use really isn't such a profound issue.
I'd imagine that your average 150k mile car will be getting worse MPG than its best, and might well have the looming bill-cloud of a new turbo / dpf / injectors hovering over it, poised to strike in much the same way.
Fortunately if the battery is degrading due to use, you've got the choice of whether to replace it or accept the lower range. If your was-250-now-200 mile range car is only used for commuting, you might not be so bothered. If your petrol car needs a new GPF and clutch, you've probably far less choice about whether or not to pay to repair it.
I'm not saying that electric cars will be without their flaws, but please let's not forget that petrol/diesel cars that are getting on a bit are also vulnerable to large bills.
There's one gotcha in your analysis; you have assumed that the petroleum arrives pre-refined. Does the refining process not consume vast quantities of electricity - or, in other words, do our existing petrol/diesel cars place an indirect burden on the grid?
Or in other words, the step 'up' to charge an electric car is offset by a step down in power requirement for the reduced oil consumption, closing the gap somewhat. Fewer new power generators needed!
As a few others have alluded, it's not that people are in need of a sustained 1gbit for hours on end for $task, it's that some limited tasks become much smoother with a higher speed connection.
Take even streaming video now at 1080p. Might only need 20mbit; but if your player wants to buffer 10 seconds of video in RAM before it begins playback, the time it spends buffering will drop to almost near-zero with a gigabit connection.
I write this from the lucky position of having Hyperoptic* 1gbit internet. I've found that I've stopped really caring about 'the internet' in terms of speed, stuff just works. If I'm sending a multi-GB file to cloud storage, it's done in barely no time at all. If I want a 15GB patch to download, it's done in the time it takes to make a cup of tea. A friend of mine languishing on 10mbit ADSL has to leave his computer on most of the day to grab the same updates in time for a gaming session in the evening.
it isn't that I need to shift 50TB of data a month, but that the data I do transfer, I don't want to wait for. I now don't have to.
(*Granted, the speed is lovely. Pity they still haven't rolled out ipv6 to me, nor can they offer a date when it'll arrive....)
Well, not strictly - I mean in the case of Dropbox where they merely used S3 for object storage, but metadata etc was on their own servers, they could have presumably moved an object and then updated the pointer to that particular object to its new home, be it a new S3 bucket, a new S3 region, a new S3-alike provider, or their own infrastructure. The delete from the original location could then take place later.
Then, once tested, you could migrate across in a number of ways - new files written to the new storage, older files moved across in the background.
Of course the migration will have a fee; the bandwidth egress charges from your old provider are going to be pretty high due to the migration!
Compiled very quickly on this i5 Macbook Pro.
I pointed it at a 5888x3912 8bit colour PNG source, with quality set to 95. It pegged one CPU core at 99%, with memory usage just shy of 5GB. 23 minutes later, I had myself a nice small JPEG. Looked perfect!
I concur with the researchers - this is worth doing where the bandwidth savings will be significant, or if your application demands a very small file size. But for general purpose use it's probably much easier not to bother. Focus the attention on the 2MB javascript file and all the work that spends doing instead!
Looking at the github page (https://github.com/google/guetzli) it is specifically an encoder for JPEGs. The decoding is trivial - if anything, it'll use less battery since there will be less data to transfer & thus the cell radio can power down sooner.
The real cost is the one-off computation to produce this smaller JPEG file. Potentially worth it for someone like the BBC to save bandwidth costs for images on their news site, less so for a low-traffic blog.
Perhaps they might be looking to set availability flags for each of their datacentres based on the local power / cooling requirements? Windy, cold day in Ireland? Get it doing all the work. A warm still evening with no solar, shift the load elsewhere.
While this might not solve all the problems it could certainly help out.
I think it's more that the market expected there to be a gold-rush style sale of new gear (running Windows 7 / 8.1 or similar) to replace old unsupported XP kit, so they filled their warehouses with lots of lovely stock to try and push on businesses and customers that were upgrading.
The upgrade-rush didn't have the scale they apparently hoped for, and now the vendors are staring forlornly at all this gear they didn't sell.
Rather than dual-sim phones, would it not be better if the networks had an easy way to report bad coverage in areas & if they were seen to react to these reports.
Just think - how many people do you know who are on a particular network because that's the only network providing good coverage in a location the user frequents - but they'd be off to a competitor if they could.
Ah, I suppose that'll never happen, will it?
Why not 3A? Well the capacity is probably measured at a tremendously slow rate of discharge. Thanks to Peukert's Law (http://en.wikipedia.org/wiki/Peukert%27s_law) the relationship between joules delivered and rate of discharge is not linear.
Therefore a 3000mAh battery might only be able to sustain a current draw of 1 Amp for half an hour* rather than the 3 hours a linear relationship might suggest.
*I made this up, did no maths at all to provide that number.
The trouble is that by attenuating the signal from the macro cell outside, your mobile phone will ramp up its TX power to overcome that same attenuation, therefore undoing your careful work.
For me it's also annoying - I live in a new build property near a city centre. Full five-bar signal outside, a flickering feeble bar indoors. A Three HomeSignal has fixed the issue there, but given the postcode they were certainly reluctant to give me one! What forced their hand was that the advisor asked to phone back on my landline midway through the call as the call was breaking up!
Had another thought about O2's capacity.. perhaps with their coverage obligation they expect to have many more masts (perhaps to ensure no building is too far away from one). As it follows that each mast will cover fewer users, this allows them to go with less spectrum. They'll spend the same, just the spending scales are tipped in the direction of O2's infrastructure, away from the OFCOM auction.
Or they're just flat broke and hoping to cheap it out all the way.
The arguments about paying double for the power and data don't really stack up in my eyes. The Three home signal in my house uses around 2-3 watts, roughly equivalent in electricity costs to one pint per year.
It uses around 2MB of data per day when idle, calls use around 0.75MB per minute (so around the same as a SIP call, give or take). SMS I can't really measure, but it'll be trivial.
I'll write off any usage due to data, but suffice it to say that any browsing I do on my phone while at home will backhaul via the ADSL, either directly via the wifi or via 3G to the homesignal and via the VPN back to Three.
So in terms of financial and bandwidth costs, it's trivial. In exchange, I get full 3G signal around the house, rather than a flickering 0-1 bar signal. The device was sent to me for no charge.. Colour me happy! Now if only my Three phone worked in other indoor areas...
Weighting it by current popularity should be fine. I suppose a comment on the software saying "You can use one of these browsers to merely search for and download any other browser you prefer" would maybe help too - after all, how many of you have "www.getfirefox.com" as the first and only link you visited using IE on a fresh windows install? Sure, Firefox will make it to the big 5 listed initially by the software, but, same concept.
Sure, you'll need access - but that was the point. Consider an iPhone being stolen, with sensitive data (presumed to be) protected by the encryption.
With a stolen phone, you can bypass its encryption and view the data anyway. The Blackberry won't give up its secrets so easily, if the encryption is enabled!
...Surely this system would therefore require nearly double the capacity on the air conditioner?
I assume that it will operate twice as hard during the night, thus eliminating the need for daytime operation of the compressors. But, surely overnight the system will be chilling for tomorrow, AND will be coping with the continued operation of the data centre. Ok, sure, loading on the centre might fall, and also the efficiency of the cooling cycle (they're far more efficient across small temperature differences) will make calculating this difficult, but it's a consideration!
To those offering that the off peak power might become on peak; surely the trick is to use smart metering and control, allowing excess power to be used whenever it's there; so that even if the night time becomes the new peak, the load could spread out a bit.
How did they spend that much?
Power consumption? Er, yeah, take a big gaming laptop, and measure the power it uses while charging its battery AND playing crysis on medium. That's the most that each socket would reasonably use.
Design? Hmm, I could knock up something graceful looking in CAD that'd do just that job.
Cost to implement? Few days labour I'd imagine.
... I'd love to see them explain where that cash went! Just out of curiosity.
For a sales team, it might be worth investing in a dual-WAN router (these can be had for just over £100), and then a second broadband connection from a totally different provider.
That way if one provider goes down, you'll only notice a reduction in speed, rather than the complete absence of an effective sales team.
Why do you enclose the "important points" with the quote marks? That just reminded me of the Lasers being described by Dr. Evil in an Austin Powers movie.
netstat -b simply lists all the open connections your computer has. As I browse El Reg, my computer must connect to its servers in order to download the web pages. Similarly, if I sign into MSN, Skype, connect to a Ventrilo server, you name it, the connection is made and is listed by the netstat command.
I don't have VM Broadband, nor do I have VIrgin Media's PC guard installed, yet this is still the case. It's expected behaviour and is not fishy. If you're thinking about Phorm analysing the websites you go to, there's no software on your PC which you can uninstall to remove that; Virgin Media watch the traffic flowing off their network further up in their network, for that.
If it weren't for Phorm, i'd like the Virgin Media method. As has been said, it's a consumer grade product; they state "unlimited" as, still, you get charged no more, no matter how much you download. Sure, it slows down at the peak times, and they do clearly define their traffic shaping policy, but fundamentally they won't cut you off altogether at a defined amount of usage.
I can see where this will go, most of the UK ISPs will start bricking it due to their lack of investment in bandwidth, just like with the iPlayer.
This is the time they've been dreading, the introduction of something legit, mass-market, legal and bandwidth intensive; just what they don't want their UNLIMITED!* subscribers to start using.
I'm the proud owner of an IBM T40 laptop, which sucks air in from.. the side, and pushes it out from.. the side! I've got some software which shows me its own temperature readings from all over the laptop (including two sensors in the battery pack itself), and when left on a bed or sofa, they're all in the low/mid 30s (unless the CPU is working hard, when that on its own increases a little). Moving it to a desk merits a temperature drop of all but a couple of degrees.
Fine? I think so, especially since the battery's warning label lists 100°c as the maximum temperature; OK having the battery warmer will shorten its lifespan, but i'm already well aware of that and happy to tolerate it.
It seems to me that bad design is the culprit here.
I thought there were two things upon which the speed of your connection hinged.
Firstly is the speed of the link between yourself and the exchange. The best "real-world" indicator of this is the line attenuation (which is often shown in your ADSL router's status page, if you've already a connection) - lower means better.
Secondly, is further up the rung - as has already been stated, many ISPs oversell the bandwidth; they assume that not everyone will be on at the same time - this is the contention ratio, often around 50:1. So they might put, say, 50 customers who've all been advised "up to 8 meg" onto a single 8mbit connection to the internet. If the ISP can't afford enough bandwidth, then things will start to slow down horribly at peak times.
The first issue is troublesome - BT are the only people likely to care, and so long as you can get around 0.4mbit (i think...) out of your line, they won't do a thing. Only possible solutions here are to move closer to an exchange (yeah, right..) or install a new line in the hope that you get a newer bit of wire! The ISP isn't able to do a thing, besides connect your wire to the equipment in the exchange and hope - no amount of shouting will make them change a thing really, nor will "I pay £20 a month i should at least get 20meg" - for ADSL, that's just not going to happen.
The second, the bandwidth issue, is however where the ISP has a shout. They've got to find the ideal trade-off between speed and cost. To give every user the full (for example) truly unlimited, 8mbit connection would cost much much more than the £24.99 a month that the "unlimited*" plans give. A very good website to check, incidentally, is:
http://www.kitz.co.uk/adsl/caps.htm
The Register - Independent news and views for the tech community. Part of Situation Publishing
Biting the hand that feeds IT © 1998–2021
