UK minister recalls two planning decisions which blocked datacenter investment The UK's deputy prime minister is set to recall two planning decisions which have held up datacenter investment in the UK. A speech by incoming finance ministers outlined how the new government would unblock the planning process to try to boost the economy. Chancellor of the Exchequer Rachel Reeves noted that the Deputy Prime …
That's not entirely wrong though - for some applications latency really does matter.
And given the lack of QoS on most networks, you have to improve the latency of everything to improve it for the traffic that actually needs low latency.
Putting the datacentre near the end points will help.
Alan Sugar will probably like it. He's gotten all bent out of shape due to people working from home, and why? Because his money is in office space. He maybe a billionaire but he made cheap products with cheap parts just at the right time.
Most of his kit was shit and many moons ago he was on Jon Ross touting the animatronic of his face, like the singing fish. Really shitty devices.
https://youtu.be/Tnh9VNVdVyI?si=iLNfAP9QO1u-H_aX
He even admitted once, not to long ago, don't start up a business, just buy property. Mainly because all his money now is in those properties. And with everyone working from home, he won't be able to sell his office space. And as always, fuck whit managers either don't see this or want to kiss his arse, despite never meeting him.
My long winded point. Alan could stop moaning about people working from home and instead change some of his buildings into data centres.
Yep. Although he's genuinally more successful than Trump and less of a cunt, I've come to realise I'm not sure he's the king of business he's been made out to be.
Lee Mack was on a chat show with him and Pamela Stephenson. Lee Mack was nervous and said Alan was being derogatory about Pamela's job as a psychologist. Something about it not being a real title or something (Lee tells this better than me on the Richard Herring podcast). Lee piped up "No Alan, its not everyone that can buy a title". Lee felt good, he hadn't said much and this was his best joke. Unfortunately he said Alan called the BBC later and demanded they cut that bit :o) and them being the spinless fucks they are, did.
And given the lack of QoS on most networks, you have to improve the latency of everything to improve it for the traffic that actually needs low latency.
Kind of. Network latency is pretty much distance, and stuff like synchronous replication works up to around 100km with modest amounts of pain. However if your network lacks QoS, then you're probably doing it wrong. So attempting to use Internet connections and VPNs when you probably shouldn't. But this is a classic problem with datacentres, ie the availability of fibre, wavelengths or GMPLS services where you can have some deterministic behaviour from the network. Diversity is also a challenge as you move away from London given there's less choice of suppliers, or their networks assume traffic will be heading to London, or take a scenic route around parts of the UK. That makes diversity rather more challenging, especially as routes will have different latencies. Good luck with that if you're trying to do load sharing or balancing and the applications are at all delay intolerant.
>However if your network lacks QoS, then you're probably doing it wrong.
My network has a transport level minimum QoS,
But it doesn't apply different QoS to different traffic types
That would (a) need the end point to tell you what the traffic is or not encrypt it (b) have an engineering cost to implement and (c) have a legal cost in fending off the inevitable net neutrality complaints.
That would (a) need the end point to tell you what the traffic is or not encrypt it (b) have an engineering cost to implement and (c) have a legal cost in fending off the inevitable net neutrality complaints.
Yep. I suspect you've been sold something that's an Internet solution and effectively is just an SLA guaranteeing some nebulous packet loss or 'up to' language. With plenty of caveats. Like you say, you can't really do QoS or CoS across the Internet, or Internet-based VPNs without falling foul of 'Net Neutrality, or just general lack of agreements to implement it across peers. Legal costs should generally only apply to ISP/SPs as they're the ones who aren't allowed to implement QoS.
Otherwise yep, although it's more a case of you telling the endpoint (or provider) how to treat any marked traffic and remembering CoS/QoS really only works when there's congestion. EF is usually different as those packets get forwarded first. Then it's back to commercial stuff and how much your provider tries to charge for EF traffic. Generally I'd recommend avoiding those because that's a provider that hasn't got enough capacity, or figured out that a rating engine, billing system, sales training and dispute resolution costs more than 'premium' EF traffic is worth.
QoS decides what traffic you are prepared to drop (if you don't have the bandwidth) or queue (if you do have the bandwidth) to allow better service for other applications
QoS doesn't alter the unloaded latency of a link which is primarily based on distance for links over long distances or routing hops for short distances - it potentially helps with loaded latency on a link (i.e. priority queuing for voice) but that that depends on how easily the traffic can be differentiated and how well applications handle dropped packets.
Having had first hand experience dealing with the numpties in the local planning office, my experience is similar to that of many folks all over the country. The planners are keen to take your money then sit on their hands for as long as possible and if you manage to get a reply, it is likely to be a copy/paste document that indicates they haven't even read the documents they requested. Bunch of useless tossers the lot of them. It is a wonder anything ever gets built here.
"Having had first hand experience dealing with the numpties in the local planning office, my experience is similar to that of many folks all over the country. The planners are keen to take your money then sit on their hands for as long as possible and if you manage to get a reply, it is likely to be a copy/paste document that indicates they haven't even read the documents they requested. Bunch of useless tossers the lot of them. It is a wonder anything ever gets built here."
It may surprise some that this is pretty close to the sentiment of the new government. I work for government, and notwithstanding that I didn't vote for the winners of this election, what I'm seeing in these early days is a real commitment to get things done, to change stuff, and to sort out long running problems like planning, housing etc. Priority number one for this government is economic growth, and it seems to me that they're not going to worry about breaking a few eggs to make the requisite omelette. I've heard a senior government minster state (in my words) that the planning system is right in the crosshairs as a major obstacle to growth.
This is against the grim economic wreckage left by the last government's incompetence so there's only so much room for manoeuvre, but I still think many be people are going to be surprised by Starmer's government, both on the left and on the right. I think they'll be surprised by the fact that this isn't a 1970's Michael Foot style attempt to emulate Cuba, they'll be surprised that quite a few of the ministers and the PM will turn to surprisingly competent at their jobs, and I think people will in due course be quite pleased with what is eventually achieved.
There's relatively few things they can do before Parliament's summer recess and that's unfortunate, but (speaking as somebody who ought to be a Tory supporter) Starmer's cabinet and approach to leading policy has the makings of a really, really good government.
>” Starmer's cabinet and approach to leading policy has the makings of a really, really good government.”
Well from the announcements made so far, it’s as conflicted as previous administrations: saying it wants to devolve more to local government whilst also making more central dictates on housing, wind farms, data centres…
So yet to see any change from the Tories….
Labour are labelling themselves as the party of change. Well they love to change their minds depending which way the wind is blowing:
Ed Miliband, 2015:
Ed Miliband slams Government over 'devastating' sudden closure of Hatfield Colliery after 99 years
https://www.mirror.co.uk/news/uk-news/ed-miliband-slams-government-over-5978358
Ed today:
Labour says it would stop Cumbria coalmine from opening
Ed Miliband vows party will seek to prevent ‘climate-destroying’ plan and if elected would deliver green jobs
https://www.theguardian.com/environment/2022/dec/08/labour-says-would-stop-cumbria-coalmine-opening-ed-miliband
Labour now going on about reversing the ban of onshore windfarms, yet before the ban it was many of their own councils deciding against them:
https://www.theguardian.com/big-energy-debate/2014/nov/28/onshore-wind-power-council-planning-approvals
Councils are not the government, even if they wear the same colours.
Councils are supposed to care more for local things that affect living people in their area than what's good for the country. Governments are supposed to consider the overall welfare of the country, while attempting to minimise damage such considerations might cause to locals.
Therefore it's completely unsurprising that a Labour council can fundamentally disagree with a Labour government. As could any council and equivalent goverment. Not just a Labour thing.
One big frustration I have with the planning around data centres is that sensible use of the waste heat isn't a requirement. A DC using 96 MW could heat a neighborhood, or can a lot of food with some heat pump concentration to get up to the required temps.
A combined Amazon datacentre and meat packing plant?
Get 2 packs of Prime Green with every new subscription! Biggest problem with using heat is the cost of getting it to any useful temperature. I guess where offices or light industrial buildings are collocated with the datacentres, the warm air could be used to heat those and reduce their heating costs. Also curious if the one with the planned lake/pond in the centre would use that as part of the cooling system. Could be a good place to put the piranhas if it's nice and warm.
I've heard about a scam called "heat pumps" where you create useful temperature out of low temperature with a COP of 3+ ... just saying
Heat pumps are a different scam that will add energy demand, much like data centres. Curious how you'd use heat pumps for district heating, or dealing with any waste heats from a datacentre though.
Curious how you'd use heat pumps for district heating, or dealing with any waste heats from a datacentre though.
For the downvoters, have a read of this and think about it-
https://en.wikipedia.org/wiki/Cold_district_heating
Cold local heating is sometimes also referred to as an anergy network. The collective term for such systems in scientific terminology is 5th generation district heating and cooling.
Which isn't science, it's marketing guff. But basically uses lukewarm water as a working fluid rather than hot water. The advertorial merrily glosses over costs and actual efficiencies. So to do anything useful it would require a new water network and subscribers to also have their own heat pumps. If they have to have heat pumps to get water to a useful temperature, why not just do that instead rather than relying on someone to try and flog you 'anergy'? And..
If the roofs of the buildings supplied are equipped with photovoltaic systems, it is also possible to obtain part of the electricity required for the heat pumps from the roof of the consumer. For example, 20 PlusEnergy houses have been built in Wüstenrot, all of which are equipped with photovoltaic systems, a solar battery and a heat storage tank for the highest possible degree of self-supply through flexible operation of the heat pump.
Thats.. an expensive additional cost on top of any costs charged by the district heating supplier. Plus installing a 'solar battery' (wtf) and a heat storage tank requires a fair amount of space, which old and newer housing stock may not have. Good'ol hot water tanks are a thing of the past thanks to the move to gas central heating and on-demand hot water. Plus if you have solar PV, or hybrid solar-thermal and space for a hot water tank, the district heating becomes pretty much redundant anyway. Run solar PV to a cheap resistive heating element in a hot water tank and you get hot water anyway. The only gain you get is if the incoming water is slightly warmer than the mains/well water temperature, but the district heating supplier will want to charge for this. And they'll have to charge a lot given the high costs of provisioning and running their services.
(This is an exercise I'm doing for a new home, and looking to make that as efficient and off-grid as possible. Not helped by wanting to make use of every erg of energy, and that also being the noise made when seeing the costs. Also me thinking stuff like "I want a forge!", and how heat for that could be captured and used. At least with datacentres the heat output is pretty constant.)
I understand district CHP are a mandatory part of new large developments in Germany. In the UK, not at all - maybe a few solar panels and a car charger or two at best. It's insanely short sighted, and hopefully will be something the new Govt will look into.
We've tried domestic district heating in the UK but it was done alongside the dash to knock down the slums and move everyone into prefab uninsulated high-rises. And the results were BAD. Iron pipes rusting through, flooding, no heating etc. Given just how shit social housing is now people are very scared of anything that didn't work in the past. They equally can't see how things might get better so we are stuck in a paralysis.
Sadly govts are only after quick fixes and housing developers want to chuck up the cheapest crap they can get away with.
There have been lots of more recent UK neighbourhood heating schemes that have all gone badly wrong too. Houses left too hot or too cold, prices going through the roof with no regulation, sudden big charges for unplanned maintenance (more than the cost of fitting a new private boiler), long period of downtime because breakdowns.
Here's a nice TV programme from 1984 which details why all the high-rise buildings thrown up in the 60's and 70's in Britain were absolute shit.
" housing developers want to chuck up the cheapest crap"
I'll qualify that and say they want to throw up the cheapest crap they can make the most money on. Family' town' houses mostly it seems rather than smaller (and more affordable) occupancy properties.
FWIW the New Home Quality Control channel on Youtube has put me right off newbuilds. But has given me some interesting words to describe slapdash building contractors.
We've tried domestic district heating in the UK but it was done alongside the dash to knock down the slums and move everyone into prefab uninsulated high-rises. And the results were BAD.
Sometimes. And sometimes it's turned out very GOOD. Like in Southampton.
It's possibly important in Southampton that the majority of residential consumers on the Southampton network (like the apartment blocks on the former Dell stadium site) were not part of the initial 1980s build out, and they were designed, built and connected ~2000-03 after the operators had gained experience heating the Civic Centre, hospital, a swimming pool, a couple of hotels and other city-centre commercial units, some of which were tricky retrofits not new-builds designed for district heating.
My understanding from a former resident of The Dell redevelopment is that it all worked rather well.
Governance structure, funding and liability for repairs/maintenance are all fairly key. It's easy to see that a "district" heating scheme involving a fat GSHP supplying a circuit to an estate of 100 houses and run by a disinterested developer/subsidiary is liable to have problems. By contrast, a big scheme like Southampton which has major stakeholders like the council, hospitals and is key to heating shopping centres like WestQuay becomes a bit too big to fail, and - with the right legal framework - costs can be weighted towards the commercial entities rather than low-income households or whatnot.
One of the few :) I would go and look at the T&P on the wellhead in the Toys'R'Us car park when walking to the station.
"Governance structure, funding and liability for repairs/maintenance are all fairly key"
Fairly? Essential!! I think the fact that it heats/cools the council building helps. They are not going to put up with faults.
That wouldn't be the one in east London where prices went up by 350%? But that isn't as bad as the one in Poole where prices went up by 567%. That might have something to do with people's reluctance to embrace these schemes.
Use it to grow food by heating greenhouses.
No! We're supposed to be countering the greenhouse effect, not encouraging them!
But slight snag would be datacentres being built around London where land prices are stellar, ie reassuringly expensive. So any produce would also be expensive, unless those products were ones normally grown under heat lamps and currently illegal. Alternatively, it'd more likely be Battersea-style housing developments with the same risks of the DC or heat provider going bust, especially if residents are protected from extortionate communal heating charges. I think there was a proposal do do district heating at a DC in Park Royal, but as is often the case with 'renewables' sca.. I mean schemes, the economics made no sense. Then again, Millibrain is back in charge of energy policy so we'll probably end up funding more of these unicorns.
Plans fo more data centres in west London and Slough have already been delayed due to lack of capacity in the local electricity distribution network, so where does the Deputy PM expect to find sufficient capacity? Assuming you had 96MW capacity available, you can power a lot of homes with EV chargers in the green belt. Just saying . . . Policy confusion and conflict ahead.
Chicken and egg. If they get planning permission to build the data centre then an electricity supplier can start arranging supply. Without the planning permission granted the electricity supplier won't spend money providing 96GW to an location that doesn't need it.
The existing lack of capacity in the area for the other DCs you mention means there is now a bigger economy-of-scale incentive to an energy supplier that maybe wasn't there before.
The existing lack of capacity in the area for the other DCs you mention means there is now a bigger economy-of-scale incentive to an energy supplier that maybe wasn't there before.
It's not that simple, or cheap, or quick. There's no energy supplier that can easily supply electricity given all the moving parts, ie actually building generating capacity and getting it to the datacentre's location. So minimum of 3 parties involved, a supply company, a generating company and transmission. Lack of transmission capacity is one of the big reasons why there's capacity constraints around London and why National Grid wanted a lot more money. Those upgrades are neither quick, nor cheap. Especially given datacentres will want diversity from both transmission and suppliers. SMRs could be collocated, depending on site or just reduce transmission costs, but aren't there yet. Old stand-bys like big diesel gensets are being regulated out of the market, as are local gas turbines. Relying on the weather and very, very large piles of batteries is both expensive, and very risky.
Not forgetting that people dislike pylons so new cabling has to go underground which is hugely expensive and very disruptive.
Indeed. I was chatting with a power guy recently about this and there's a bit of an issue at the moment. Big transformers and switch gear for substations has never really been an off-the-shelf product, so usually long lead times. Those lead times have got a lot longer given demand to rebuild Ukraine's electricity grid. Which is a apparently more complicated given a lot of their network uses a Soviet-style system of HV steps to the rest of Europe. But they have a lot of transformers and other kit that needs replacing at the moment.
happens in the shires too.
The plan to bring offshore wind energy through East Anglia has some of the locals much aggrieved. The options are a line of pylons across the countryside (it already has plenty of pylons) or running an underwater cable round the coast. Obviously one option is much more expensive than the other.
saw one commentator on local website saying they thought putting up pylons would reduce the agricultural output of the land [significantly]...
The existing lack of capacity in the area for the other DCs you mention means there is now a bigger economy-of-scale incentive to an energy supplier that maybe wasn't there before.
It's not that simple, or cheap, or quick. There's no energy supplier that can easily supply electricity given all the moving parts, ie actually building generating capacity and getting it to the datacentre's location. So minimum of 3 parties involved, a supply company, a generating company and transmission. Lack of transmission capacity is one of the big reasons why there's capacity constraints around London and why National Grid wanted a lot more money. Those upgrades are neither quick, nor cheap. Especially given datacentres will want diversity from both transmission and suppliers. SMRs could be collocated, depending on site or just reduce transmission costs, but aren't there yet. Old stand-bys like big diesel gensets are being regulated out of the market, as are local gas turbines. Relying on the weather and very, very large piles of batteries is both expensive, and very risky.
Firstly there are suppliers such as SSE with generating and distribution operator licenses. National level transmission is the National Grid, of course. So I'm sure they'd be delighted to bid for provision of 147GW to the proposed DCs. Whether their price and timescale is acceptable is a different issue!
But, and perhaps more to the real point you're making, a lack of capacity will always remain a lack of capacity until someone funds it somehow. National Grid / Offgen make predictions of demand and therefore where new capacity will be required but these can't allow for DCs not yet proposed and proposers of DCs can't (easily) work out where capacity will be when they come to start operations. Again a chicken and egg situation.
Resolving that chicken and egg scenario is not simple but denying planning permission is certain to never achieve it.
Firstly there are suppliers such as SSE with generating and distribution operator licenses. National level transmission is the National Grid, of course. So I'm sure they'd be delighted to bid for provision of 147GW to the proposed DCs. Whether their price and timescale is acceptable is a different issue!
Well.. there is no catch-22 in this case. 147GW is basically basically 50 Hinkley C's at around £1.5tn. Cabling not included. Or an infinite number of windmills and most of the UK covered in batteries. But this is why there's no catch-22, if the free rider problem is solved. Want 147GW capacity? No problem, just cough up £2tn and deliveries will start in about a decade.
Not going to happen.
But it's probably something that should happen given it's the DCs generating the demand. Want power? Pay for power. Don't try socialising those costs. Especially as a lof of DCs are pretty speculative investments, sometimes working on the assumption that the DCs tenants can figure out a way to actually make money from 'AI' before their funding runs out. Or technology shifts happen. So Telehouse Metro used to be the Stock Exchange's computer centre, built to host their mainframes. Those went the way of the dinosaur, so the LSE was left with space it didn't need, and flogged the building off. The AI bubble might (will) burst, and speculators jumping on the DC bandwagon will get burned.
All part of the fun. DC builder won't get finance to pay for a 25yr PPA for X MW/GW, but that's what they'd need to do if generators want finance to build new generating capacity. Banks and finance types have seen bubbles before and want security.. Which gets even harder when tenants might want 1, 5, 10yr deals with plenty of break clauses.. Which is also why WeWork didn't work.
So, having pledged to tear up the London green belt and build lots of new HOMES that we supposedly need (despite the fact that some ex-industrial northern towns have entire streets of houses standing empty), their top priority, on their first day in office, is to revive two dead DATACENTER planning applications from Greystoke Land? Out of all the thousands of refused planning applications, those two just happened to be on the top of the pile, and the developer doesn't even have to go through the proper legal process of making a new application? Doesn't that sound a little bit fishy? It's almost as if Greystoke Land has handed them a big brown envelope. That's blatant corruption on a level we haven't seen since Boris Johnson's government. The next few years are going to be interesting.
The real reason for building on Green Belt is that land around London is super expensive. But the Green Belt land is cheap (relatively) because you are not allowed to build on it.
When land is removed from the Green Belt and allocated for development, the people who own that land suddenly become VASTLY wealthy.
That is the only reason why property speculators buy Green Belt land and then lobby the government to be allowed to build on it. There is no actual NEED to build on Green Belt land.
There are a few exemptions for building on green belt for services, but if said services are closed then the owner of the land is meant to return the land to it's original state. So the issue in your example is actually that the government is not using it's existing enforcement powers to get land owners to clean up.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
