Arcing faults tend to draw less current than a bolted fault. The GP's point was that the current can be adequate to start a fire, but not high enough to trip the transformer primary fuse, and is spot-on. Primary fuse protection can be as much as 6x full load current, although this wouldn't normally be done on residential services.
Ground Source Heat Pump. Earth temperature is constant, equal to the average annual local temperature. It is possible to build a passive ground source heat pump, but not easy.
As for the triple-glazing... well, have less window area, provide heat shades for night time in the winter, focus on radiant heating and cooling.
Passive energy efficiency is the best way to go, but there is no excuse for using electric radiant heat in a cold climate, or air conditioning systems that aren't optimized for local climate.
You are focusing on the tradition way to build a telecom infrastructure. What unregulated competition encourages is "lean and mean" telecom providers that can survive by creating a cost-effective infrastructure... even if it requires them to mountn $10 hubs on telephone poles.
The telephone (from the user interface perspective) did not change much over the course of 100 years. How can you use that same mindset in an era where the customer demands on the infrastructure double every couple of years?
The incentive to unregulated markets is that if your competitor makes a bad decision, or lacks vision, you can better compete in the market-- either a specific market segment or overall.
A dolphin kick is something that divers occasionally do with normal fins to vary muscle usage and avoid fatigue and cramps. So many of us are somewhat familiar with the general style. The problem with this style is that it is quite limited with respect to maneuverability. Divers often use their legs/fins asymmetrically or at odd angles. This far more useful than going fast.
I was thinking the same thing... the versatility of a normal rigid fin seems nicer than a "high speed" fin. Haven't tried doing a dolphin kick with a split fin, but not sure how you might be able to do something along the lines of sculling with this would seem like a challenge.
Not all diving is in wrecks... but in a wreck you want something that doesn't create too much vertical turbulence, which will stir up silt. Curious how this would do there...
$500 is perfectly reasonable, agreed. But the underwater scooters are hard on air consumption (you have to hold on tight and maintain leg stiffness for proper control). If the super fin works well both for high speed and low speed motion, it sounds like a winner.
Personally, I'll stick with my Quattros. Scuba diving is all about moving as little as possible... but having the extra power when you have to fight a current.
Also, there is only so much one can cut the energy use, and thus that cost down, and still get the business of the government done. And the improvements in efficiency will require hardware, software, and personnel which have their own costs. Eventually you will hit a point where there is no longer a return on investment to make it worthwhile.
The same thing was said for many other things over the years; lighting pops to mind. Offices used to consume about 3 watts per square foot of office area in the 70's. In the 80's, they got it down to 2, and in the 90's around 1.5. Today, with a bit of effort, you can get it under a half a watt per square foot. With automatic controls factored in, the daily power consumption has dropped by a factor of 6-7 on average over 35 years.
Interestingly, offices consume roughly the same amount of power today as 35 years ago. If not for lighting energy savings, many of the computer innovations over the last 20 years would not have been practical, since buildings would not have enough power and cooling capacity to support the increased loads.
Saving 50% of the total energy consumption of a PC today might help usher in the next technological innovation. Don't be too quick to poo-poo things that don't sound cool on the surface. Their just might be a market for piece-work server virtualization in a few years...
Except for mom-and-pop customers, all large facilities (365 Main, Savvis, Equinix) charge clients directly or indirectly for power consumption. They might price the cost in ways to obscure it, but if you work it hard enough you can get it down to $/SF plus $/kW. They will often give you a price break for redundant circuits over normal circuits if you are persistent enough.
The problem with the equation as you suggest is that installed capacity is more expensive than consumption-- The lifetime cost of the infrastructure is about $0.08/kW/hour (assuming linear depreciation), and it is normally well under 60% utilization. When the equipment is installed, the maintenance and operation overhead is based on that installed capacity, not end-user utilization.
It's great to introduce government mandates into things (it's the only way harmonics and power factor problems for computer equipment got solved), but the bigger value of this study is in helping people understand just how much power is wasted (power supply losses as an example) in a data center.
While I agree that avoided costs are important to consider, the problem with that argument is that you need to understand all the risks you face, the likelihood of occurance, and how changes to a design might impact other issues.
Two simple examples are the changes to the building code after the Northridge earthquake, and designing for floor vibration in a high rise. In the second case, a 100 storey building that needs a two inch deeper structure to reduce vibrations to a level where optical microscopes work would likely force two floors to be eliminated. Why design for that, unless there is a specific requirement for it? It isn't prudent engineering!
For the earthquake designs, what do you do if you have a perfectly good building that doesn't meet current codes? California forced upgrades on all of the hospitals, but not office buildings. Do you spend 20% of the building replacement cost to retrofit a 10 year old building with a likely life of 30 years against a 10% chance of an earthquake that would red-tag the building occurring in the next 50 years?
It's great to prevent loss of life, but you have to draw the line somewhere.
For towing, why doesn't someone come up with a (hybrid) powered trailer? There are plenty of needs for something with hauling capacity, but why should it be there all th time? Pickup trucks make practical sense for cargo flexibility, but that flexibility isn't required for every trip. Construction companies trying to go green should encourage more flexible options for their crews.
Plenty of applications suffer from the high bed height of a pickup.
That sounds workable; just use a thumbdrive as an intermediary for file transfers when you don't have access to the physical network. Yes, the thumbdrive is still stuck in the compromise FS (fat), but since it is just used as a transfer medium, little information is lost. Not perfect, but better for archival.
NFPA 110 only applies to life safety systems-- specifically human egress. For that, I think the SF 365 main has an extra generator (that's how some of the other ones are set up), or they could just put battery packs in the lights.
For most generators for the biggest banks, we specify redundant starting systems, but what you really end up getting isn't a 2N system (on a CAT engine for sure), unless you go with compressed air backup starting. Compressed air gets pretty tricky when you need to meet NFPA 110, or when you are running six engines in parallel.
The bigger issue will end up being fallout from co-lo in general, for companies that have either customer or UL requirements for backup provisions.
All told though, I think it is best to go with half rotary and half static UPS systems in a 2N arrangement. Rotary works great for the blips that you see most often, and the batteries work well for longer discharge times.
It takes Diesel a few years to go bad. That site has fuel polishing systems to prevent that. Because of earthquake risk, they contractually are obliged to have 24-48 hours of backup fuel with many of their clients.
They have the HiTec rotary UPSs in all their facilities, which link a generator to a flywheel UPS. It's stupid to not have backup fuel for that type of system; you can only run for 13 seconds before the load crashes.
It is possible that they got a number of small hits and the generators failed to re-start after a few. Good procedures are to stay on generator until utility stabilizes if you have more than one "hit."
The only alternative to a good frequency available nationwide is software-defined radios that can operate on a wide range of frequencies. The problem with this approach is that it drives up cost and is not power efficient.
If you think AAPL has a high PE, check out RIMM. They are pushing 100 now if memory serves me. MSFT and IBM are in single digit growth. Apple should hit about 30% for the just finished quarter, and likely 40% for the current quarter.
Relative humidity isn't a problem if you don't bring in any outside air-- just have the building vapor sealed and maybe insulated a bit to help trap the heat inside. A few 100% glycol lines to the outside and you have a nice free cooling system.
Of course, the cost of that OC192 line might slightly offset the electricity savings, but that's just a detail. Generating power for the servers might be interesting as well...
Density has a number of benefits-- especially in reducing your network and power interconnect costs. We build data centers that are about as big as a big-box retail store, and the average power draw is around 5kW per cabinet, or 125W/SF raised floor. Your 100,000 square foot retail space would tend to be half raised floor, so you have a total of about 6.2 MW UPS.
The real problem for data centers is that some equipment works much better packed close together. Usually, it's only 20% or so, but you have to figure out solutions for this type of equipment.
The most interesting strategies for data center cooling today are using air side free cooling. There are plenty of challenges, and it only works with certain combinations of local climate and building design, but it is another area that benefits from high density-- being able to exhaust 110F air from your cabinets directly to the outside rather than trying to cool it back down to 55F at the CRAC units makes a lot of sense.
(As for converting a big box retail building to a data center... you might be able to put in 100kW of computer load and just run the air conditioning at night as suggested. If you pay $0.50/SF, that would be about $50k/month in rent. Rent in a co-lo for the same power density would be about $26k per month. In either scenario, you need to add power and UPS to the equation for the total picture.)
Why the XDV/XDO and not the XDF? And looking at the XDF, which should be a better match for Dell (14kW max?!), Lennart isn't keeping up with Rittal. The XDF is a water cooled or compressor based high density cabinet.
Anybody that tries to buy this stuff from Dell and just install it like a server is in for quite some fun!
There is visual indication on the shift key-- the shift "button" glows. Looks like they didn't change text styles on the keyboard to maintain readability.
The keyboard and landscape orientation options do need more work, but it is worth the money compared to a blackberry or treo.
Over time, Apple had better open the thing up some. The calculator is a pathetic joke, and there are far too many apps that just a web interface won't work. Lack of IMAPS was pretty lame; looks like the POP isn't secured either, but I haven't done packet analysis on it. Ultimately, these are disappointments, but not deal breakers, at least for me.
The problem is power consumption of the 3G chipsets. It's too high to give the battery time in the form factor Apple wanted. Had they gone with 3G, they would have reduced the performance for the majority of users in terms of battery time, so that a few users can have 3G speeds between home and office.
Hopefully, AT&T will get a massive deployment of picocells in areas with extra need going to ease the network burden. Apparently AT&T has done a lot to open up extra slots on their EDGE network that should help ease congestion some.
When it hits Europe, 3G is a given. It just doesn't make design sense in the US at this time.
Actually, a brownout is generally a loss of frequency due to system overloading. In the good old days of resistive loads, letting the voltage sag 10% would drop the power consumption by 20%
With inverse-impedance equipment being the norm today (switch mode power supplies, electronic ballasts, and VFDs), a brown-out is much more likely to create localized outages as individual feeders become overloaded.
But, what they are most likely really referring to is aggressively scheduled maintenance to allow for upgrades. "Non-critical" power is shut down to allow for upgrades to normal sources, and less critical loads are shed to assist in upgrading backup supplies.
Brownout just sounds better when they haven't lost all power...
I kinda fear this outside our data center. Especially when the machines therein get on the "long in the tooth" side, and we've decommissioned every application in the thing but one.
Remember IBM's "lego" data center concept-- little boxes, and it grows out as time goes on, generally moving across your data center floor over the years? Same thing with this type of solution-- buy one a year, and migrate apps as time requires to new containers. It really is more of a portable data center (sans infrastructure) than a singular processing unit.
I think the main point was that Apple charges for updates within an OS line, comparable to the Service Packs for Windows which Microsoft gives to its customers for free (on the other hand, they're less frequent).
The problem with that statement is that a service pack offers no additional functionality. The OS X updates do add significant functionality that really does a lot to make them worthwhile.
As to price, I agree that it is a bit high in a frequency standpoint. However, from a TVM perspective, the value is clearly there.
I'm pretty happy to pay for updates, at least every other time.
Slashdot Mirror
Arcing faults tend to draw less current than a bolted fault. The GP's point was that the current can be adequate to start a fire, but not high enough to trip the transformer primary fuse, and is spot-on. Primary fuse protection can be as much as 6x full load current, although this wouldn't normally be done on residential services.
Ground Source Heat Pump. Earth temperature is constant, equal to the average annual local temperature. It is possible to build a passive ground source heat pump, but not easy.
As for the triple-glazing... well, have less window area, provide heat shades for night time in the winter, focus on radiant heating and cooling.
Passive energy efficiency is the best way to go, but there is no excuse for using electric radiant heat in a cold climate, or air conditioning systems that aren't optimized for local climate.
You are focusing on the tradition way to build a telecom infrastructure. What unregulated competition encourages is "lean and mean" telecom providers that can survive by creating a cost-effective infrastructure... even if it requires them to mountn $10 hubs on telephone poles.
The telephone (from the user interface perspective) did not change much over the course of 100 years. How can you use that same mindset in an era where the customer demands on the infrastructure double every couple of years?
The incentive to unregulated markets is that if your competitor makes a bad decision, or lacks vision, you can better compete in the market-- either a specific market segment or overall.
I was thinking the same thing... the versatility of a normal rigid fin seems nicer than a "high speed" fin. Haven't tried doing a dolphin kick with a split fin, but not sure how you might be able to do something along the lines of sculling with this would seem like a challenge.
Not all diving is in wrecks... but in a wreck you want something that doesn't create too much vertical turbulence, which will stir up silt. Curious how this would do there...
$500 is perfectly reasonable, agreed. But the underwater scooters are hard on air consumption (you have to hold on tight and maintain leg stiffness for proper control). If the super fin works well both for high speed and low speed motion, it sounds like a winner.
Personally, I'll stick with my Quattros. Scuba diving is all about moving as little as possible... but having the extra power when you have to fight a current.
The same thing was said for many other things over the years; lighting pops to mind. Offices used to consume about 3 watts per square foot of office area in the 70's. In the 80's, they got it down to 2, and in the 90's around 1.5. Today, with a bit of effort, you can get it under a half a watt per square foot. With automatic controls factored in, the daily power consumption has dropped by a factor of 6-7 on average over 35 years.
Interestingly, offices consume roughly the same amount of power today as 35 years ago. If not for lighting energy savings, many of the computer innovations over the last 20 years would not have been practical, since buildings would not have enough power and cooling capacity to support the increased loads.
Saving 50% of the total energy consumption of a PC today might help usher in the next technological innovation. Don't be too quick to poo-poo things that don't sound cool on the surface. Their just might be a market for piece-work server virtualization in a few years...
Except for mom-and-pop customers, all large facilities (365 Main, Savvis, Equinix) charge clients directly or indirectly for power consumption. They might price the cost in ways to obscure it, but if you work it hard enough you can get it down to $/SF plus $/kW. They will often give you a price break for redundant circuits over normal circuits if you are persistent enough.
The problem with the equation as you suggest is that installed capacity is more expensive than consumption-- The lifetime cost of the infrastructure is about $0.08/kW/hour (assuming linear depreciation), and it is normally well under 60% utilization. When the equipment is installed, the maintenance and operation overhead is based on that installed capacity, not end-user utilization.
It's great to introduce government mandates into things (it's the only way harmonics and power factor problems for computer equipment got solved), but the bigger value of this study is in helping people understand just how much power is wasted (power supply losses as an example) in a data center.
While I agree that avoided costs are important to consider, the problem with that argument is that you need to understand all the risks you face, the likelihood of occurance, and how changes to a design might impact other issues.
Two simple examples are the changes to the building code after the Northridge earthquake, and designing for floor vibration in a high rise. In the second case, a 100 storey building that needs a two inch deeper structure to reduce vibrations to a level where optical microscopes work would likely force two floors to be eliminated. Why design for that, unless there is a specific requirement for it? It isn't prudent engineering!
For the earthquake designs, what do you do if you have a perfectly good building that doesn't meet current codes? California forced upgrades on all of the hospitals, but not office buildings. Do you spend 20% of the building replacement cost to retrofit a 10 year old building with a likely life of 30 years against a 10% chance of an earthquake that would red-tag the building occurring in the next 50 years?
It's great to prevent loss of life, but you have to draw the line somewhere.
For towing, why doesn't someone come up with a (hybrid) powered trailer? There are plenty of needs for something with hauling capacity, but why should it be there all th time? Pickup trucks make practical sense for cargo flexibility, but that flexibility isn't required for every trip. Construction companies trying to go green should encourage more flexible options for their crews.
Plenty of applications suffer from the high bed height of a pickup.
That sounds workable; just use a thumbdrive as an intermediary for file transfers when you don't have access to the physical network. Yes, the thumbdrive is still stuck in the compromise FS (fat), but since it is just used as a transfer medium, little information is lost. Not perfect, but better for archival.
NFPA 110 only applies to life safety systems-- specifically human egress. For that, I think the SF 365 main has an extra generator (that's how some of the other ones are set up), or they could just put battery packs in the lights.
For most generators for the biggest banks, we specify redundant starting systems, but what you really end up getting isn't a 2N system (on a CAT engine for sure), unless you go with compressed air backup starting. Compressed air gets pretty tricky when you need to meet NFPA 110, or when you are running six engines in parallel.
The bigger issue will end up being fallout from co-lo in general, for companies that have either customer or UL requirements for backup provisions.
All told though, I think it is best to go with half rotary and half static UPS systems in a 2N arrangement. Rotary works great for the blips that you see most often, and the batteries work well for longer discharge times.
It takes Diesel a few years to go bad. That site has fuel polishing systems to prevent that. Because of earthquake risk, they contractually are obliged to have 24-48 hours of backup fuel with many of their clients.
They have the HiTec rotary UPSs in all their facilities, which link a generator to a flywheel UPS. It's stupid to not have backup fuel for that type of system; you can only run for 13 seconds before the load crashes.
It is possible that they got a number of small hits and the generators failed to re-start after a few. Good procedures are to stay on generator until utility stabilizes if you have more than one "hit."
Be interesting to find out what happened.
Well put.
The only alternative to a good frequency available nationwide is software-defined radios that can operate on a wide range of frequencies. The problem with this approach is that it drives up cost and is not power efficient.
If you think AAPL has a high PE, check out RIMM. They are pushing 100 now if memory serves me. MSFT and IBM are in single digit growth. Apple should hit about 30% for the just finished quarter, and likely 40% for the current quarter.
RIM on the other hand is in for some troubles...
Relative humidity isn't a problem if you don't bring in any outside air-- just have the building vapor sealed and maybe insulated a bit to help trap the heat inside. A few 100% glycol lines to the outside and you have a nice free cooling system.
Of course, the cost of that OC192 line might slightly offset the electricity savings, but that's just a detail. Generating power for the servers might be interesting as well...
Density has a number of benefits-- especially in reducing your network and power interconnect costs. We build data centers that are about as big as a big-box retail store, and the average power draw is around 5kW per cabinet, or 125W/SF raised floor. Your 100,000 square foot retail space would tend to be half raised floor, so you have a total of about 6.2 MW UPS.
The real problem for data centers is that some equipment works much better packed close together. Usually, it's only 20% or so, but you have to figure out solutions for this type of equipment.
The most interesting strategies for data center cooling today are using air side free cooling. There are plenty of challenges, and it only works with certain combinations of local climate and building design, but it is another area that benefits from high density-- being able to exhaust 110F air from your cabinets directly to the outside rather than trying to cool it back down to 55F at the CRAC units makes a lot of sense.
(As for converting a big box retail building to a data center... you might be able to put in 100kW of computer load and just run the air conditioning at night as suggested. If you pay $0.50/SF, that would be about $50k/month in rent. Rent in a co-lo for the same power density would be about $26k per month. In either scenario, you need to add power and UPS to the equation for the total picture.)
Why the XDV/XDO and not the XDF? And looking at the XDF, which should be a better match for Dell (14kW max?!), Lennart isn't keeping up with Rittal. The XDF is a water cooled or compressor based high density cabinet.
Anybody that tries to buy this stuff from Dell and just install it like a server is in for quite some fun!
There is visual indication on the shift key-- the shift "button" glows. Looks like they didn't change text styles on the keyboard to maintain readability.
The keyboard and landscape orientation options do need more work, but it is worth the money compared to a blackberry or treo.
Over time, Apple had better open the thing up some. The calculator is a pathetic joke, and there are far too many apps that just a web interface won't work. Lack of IMAPS was pretty lame; looks like the POP isn't secured either, but I haven't done packet analysis on it. Ultimately, these are disappointments, but not deal breakers, at least for me.
The problem is power consumption of the 3G chipsets. It's too high to give the battery time in the form factor Apple wanted. Had they gone with 3G, they would have reduced the performance for the majority of users in terms of battery time, so that a few users can have 3G speeds between home and office.
Hopefully, AT&T will get a massive deployment of picocells in areas with extra need going to ease the network burden. Apparently AT&T has done a lot to open up extra slots on their EDGE network that should help ease congestion some.
When it hits Europe, 3G is a given. It just doesn't make design sense in the US at this time.
Actually, a brownout is generally a loss of frequency due to system overloading. In the good old days of resistive loads, letting the voltage sag 10% would drop the power consumption by 20%
With inverse-impedance equipment being the norm today (switch mode power supplies, electronic ballasts, and VFDs), a brown-out is much more likely to create localized outages as individual feeders become overloaded.
But, what they are most likely really referring to is aggressively scheduled maintenance to allow for upgrades. "Non-critical" power is shut down to allow for upgrades to normal sources, and less critical loads are shed to assist in upgrading backup supplies.
Brownout just sounds better when they haven't lost all power...
Remember IBM's "lego" data center concept-- little boxes, and it grows out as time goes on, generally moving across your data center floor over the years? Same thing with this type of solution-- buy one a year, and migrate apps as time requires to new containers. It really is more of a portable data center (sans infrastructure) than a singular processing unit.
So it only holds 6 racks? Sounds like a neat set-up, but no more space-efficient than a normal data center.
I guess they have sold at least one more than APC's "Data Center on Demand" with the Stanford purchase.
We had a concept that could hold a bit more equipment, but this seems to be pretty hassle-free.
The problem with that statement is that a service pack offers no additional functionality. The OS X updates do add significant functionality that really does a lot to make them worthwhile.
As to price, I agree that it is a bit high in a frequency standpoint. However, from a TVM perspective, the value is clearly there.
I'm pretty happy to pay for updates, at least every other time.
With customer service, stating such and being given either a significant credit on service or an "out" on the contract.
Both parties have to live up to their obligations in a contract. If they aren't, take them to small claims court (or something).