it does provided a marked improvement in performance if you're actually using your pipe.
If you're running 1Gbps Ethernet over Cat6, you will have the exact same performance as running 1Gbps Ethernet over Cat5e (or anything else for that matter). 1Gbps is 1Gbps. The only thing that will make a difference is if there are transmission errors, which I've only seen with really bad wiring jobs.
It may have escaped your notice, but burning natural gas releases energy (and carbon).
It may have escaped your notice, but this discussion is about the production and consumption of electricity. Saying that Canadians use more energy through wood burning stoves, or natural gas, wouldn't be particularly relevant now would it.
Remember to play nice on the internet, lots of other people are using it too.
At least most places that get 115F are dry enough to use a swamp cooler (a device I'd never even heard of until moving to California). If it's 90F and 98% humidity you will not cool down. Not even at night. You'll be laying in your bed, in a pool of your own sweat (because it can't evaporate) wishing for air conditioning so you could sleep.
Americans in hot regions do get spoiled by their air conditioners, but people in cooler climates really do have no idea what it's like to need one to function.
Someone pointed to the light rail article on Wikipedia which says that construction costs are higher, so I imagine heavy rail would be even higher. But the theoretical operational costs per user would be lower for light rail versus asphalt (dependent on lots of people taking the rail).
Right of way purchases out on the country are cheap, so that's not much of an issue, but right of way in the city can be insanely expensive so I guess that could easily swing it the other way.
What I'd really like to see are construction and operational costs of average rails in Europe as those are really the best case anyone could hope for. (Let's face it, rail projects in the US have been terribly managed and are never efficient.) And then some figures for asphalt and cement in the US, as those are figures we know would be accurate.
I think most of the places we saw them they were spelled pretty similar to "pharmacy". It's a shame I didn't take a picture of the text so I would know for sure.
I thought we were talking about heavy rails, not light rails. Is the article talking about light rails?
Still, the article is interesting. The cost per mile is significantly higher for light rails than freeways, which is what I thought. Unfortunately it gives most of it's monetary figures based on theoretical and best case scenarios passengers per hour rather than typical usage. And even the numbers seem a little strange. The cited study says the showcase Calgary manages a cost of $2400 per weekday passenger (versus the more typical $12-30k) but that the cost is a mere $0.23 per ride. I never saw a ride anywhere near that cheap, even in Europe where they have passenger trains down to a science.
Still, it strongly illustrates the point that the viability of these systems is about moving the maximum number of people at once between two points, ala a freight train. Unfortunately most of the time you are nowhere near capacity, which completely throws off any idealistic cost quotes.
Any citations for power useage of trains versus other systems?
I see how you misunderstood what I said. Give the person a $20,000 car. Then use $14,000 to pay for all vehicle gas and maintenance. This would still be $1,000/year cheaper than the quoted $35,000/year, and far more useful.
This shouldn't be considered practical, which demonstrates just how impractical the situation InsaneProcessor was describing.
Do you have a citation for this? I thought it was the other way around, and some quick Googling didn't pull anything up for me.
That's not a fair comparison. You're going through one of the most expensive construction projects ever built (20 mile undersea tunnel), and you can't go by road anyway.
But you can go by plane, which is further, which I mentioned. You can also take a ferry, which is cheaper, and was also recommended to me by some european friends. But because of our schedule neither was practical, so we paid more.
Did you spend any time in any of the cities you visited? You can fit 500+ people on one train arriving in London at 8.30am. Would 500 more cars be better?
That's a good question. If they are coming from main Europe, planes would be more flexible and cheaper. There are a lot of factors to be considered, and I just don't know them all. I just know that trains aren't very flexible in where they go to/from, cost a lot energy to move quickly from one point to another, and seem to cost a fair amount of money.
I'm really hoping for a large scale personal rapid transit system to be more viable, but I doubt it will ever happen.
I would put the comfort of sleeping in my reclined car seat to be about the same or a little higher than one of the economic "6 people to a cabin" beds. And I can sit up in my car.
But it doesn't matter because there are specific reasons for taking a night train with a bed that driving a car can't compare with. I was just noting that I didn't find the sleeping platforms much of a "proper bed".
AFAIK, cost of building a mile of rail is far more expensive than a mile of asphalt, though there is less maintenance for the rail. (Life cost of concrete roads is much lower, but initial investment is high.) Rail is also not nearly as flexible in routing as roads. Those 4+ lage highways cost a lot, but rail systems are a completely different set of functionality.
I'd love to see the development of an extensive thin rail system using 3 passenger cars that are individually routeable. I think it'd be a happy median between the two.
IF they decide this - as most municipalities in Europe have - then the answer to your question is "eh, no big deal, I can take the bus to the gym/grocery store."
I just got back from Europe and was surprised at the sheer number of "pharmacies" (marked with a green plus neon sign). These were easy to see, but if you looked closely were about as common as little food markets. Standing in front of my hotel in Paris could see three on the same street block as me.
People in Europe don't need to take the bus to the grocery store, and they don't buy groceries a week or so at a time. They have food markets a 100 feet from where they live, and they go there daily. Basic food distribution and practices are completely different in Europe than the US.
I did the same trip this past December. Prices were high, beds were terrible, with 6 beds in a room, dirty bathroom down at the car, and a German conductor that thought if she just talked louder in German we might understand the complex instructions she was giving us. And standing around in the unheated Paris train station for an hour in 20F temperatures was unpleasant.
But that's probably just the difference between economy and first class night train.
Cheap asphalt with far more flexibility in where to go.
I'm from central Texas, and I recently spent a few weeks traveling by train in Europe. I would say that most of the train rides were far more expensive than making the same trip by car or plane. For instance, taking the Chunnel from London to Belgium was hundreds of dollars. But flying from Glasgow on Ryan Air would have been half the cost. The problem was that it wasn't really convenient to get to/from the airports. And renting a car wasn't practical because international car travel was expensive, especially when most of your destinations aren't very car friendly.
Big trains are great for slowly moving lots of heavy freight around, but it seems a bit lacking for moving lots of people around quickly.
Funny you should mention Adobe, we actually use Adobe Creative Suite Design Premium at work on a Terminal Server. We use a volume license from Adobe that has provisions specifically for use on terminal servers.
I would rank the experience of using Photoshop on it as mediocre. Everything works fine, except that there is a certain amount of latency that makes it annoying. Users using InDesign have never complained.
There were cost savings with this method. The licenses are more expensive, but we only needed 5 licenses for about 50 people as they generally aren't using the applications constantly. If we were purchasing for individual machines, we would have purchased only the applications they use instead of the entire suite, but this provides all applications to all users in the organization at all times. However if you have a graphic design artist that spends all day every day in Photoshop, it would make much more sense to just install it locally.
* We are actually using Citrix on our Terminal Servers, which I think is the source of the latency issues. Citrix is wonderful for people accessing their applications from home or while driving in a car, but it's use internally on a Gigabit LAN seems to add too much latency and unnecessary compression.
I think (not certain) that the reason you should prefix GNU to Linux is that Linux typically includes the entire GNU toolset to form an OS. The Linux kernel by itself is just a kernel. The Linux kernel with the GNU toolset is an actual OS, which is what most people are talking about. A program compiled with GCC though is just a program compiled with GCC.
I suspect they may have been doing things 'wrong', ie the "everything in one giant database" approach. It is impossible to tune a database to deal with all sizes of data efficiently, so the easiest thing to do is break the bits apart.
Have a database that stores the various important parts of the email header viewed by users. Extract the text from the emails and attachments and store it in a different database for performing text searches. Finally, store the full email header and email body in two different databases (or just the file system), indexed by a hash on their contents (this gets you single instance storage for when an email is sent to several people).
Doing this you can use different database systems on different servers that are carefully tuned to their specific workloads.
I recently had to build an online archive to display years worth of emails stored in PST files. I couldn't find anything simple so I just imported the emails into an IMAP account, and then wrote some PHP that queried and broke up the messages into pieces for storage. Granted, it's not serving gigabytes of emails to thousands of users simultaneously, but it's fast. It really makes me wonder how people get this wrong.
*I broke up the emails into their individual MIME attachments and stored those individually after compressing them with GZIP. This meant that broken emails might end up with missing pieces, but in my testing email clients never showed the broken pieces anyway. Also, most people just want the important parts of the header and the text, so the various MIME attachments are rarely retrieved so compressing them made a lot of sense.
Well, that's what I meant to imply. You're paying enough money that you don't debug the product, the vendor does. It's that way with most products at very large corporations.
I'm not sure I follow what you're talking about. What kind of "app" would one build on the Exchange platform? As far as I knew, it was a database for storing emails, contacts, and calendar information.
A properly set up Barracuda spam filter will let users log in and see any emails to their address that have been blocked. You can then deliver the mail, and set up whitelisting options.
I would rank our experience with Barracuda as "excellent". Your mileage may vary though as blocking conditions can depend a lot on how the administrator sets it up.
Slashdot Mirror
it does provided a marked improvement in performance if you're actually using your pipe.
If you're running 1Gbps Ethernet over Cat6, you will have the exact same performance as running 1Gbps Ethernet over Cat5e (or anything else for that matter). 1Gbps is 1Gbps. The only thing that will make a difference is if there are transmission errors, which I've only seen with really bad wiring jobs.
It may have escaped your notice, but burning natural gas releases energy (and carbon).
It may have escaped your notice, but this discussion is about the production and consumption of electricity. Saying that Canadians use more energy through wood burning stoves, or natural gas, wouldn't be particularly relevant now would it.
Remember to play nice on the internet, lots of other people are using it too.
Is natural gas heating not common in Canada?
At least most places that get 115F are dry enough to use a swamp cooler (a device I'd never even heard of until moving to California). If it's 90F and 98% humidity you will not cool down. Not even at night. You'll be laying in your bed, in a pool of your own sweat (because it can't evaporate) wishing for air conditioning so you could sleep.
Americans in hot regions do get spoiled by their air conditioners, but people in cooler climates really do have no idea what it's like to need one to function.
Someone pointed to the light rail article on Wikipedia which says that construction costs are higher, so I imagine heavy rail would be even higher. But the theoretical operational costs per user would be lower for light rail versus asphalt (dependent on lots of people taking the rail).
Right of way purchases out on the country are cheap, so that's not much of an issue, but right of way in the city can be insanely expensive so I guess that could easily swing it the other way.
What I'd really like to see are construction and operational costs of average rails in Europe as those are really the best case anyone could hope for. (Let's face it, rail projects in the US have been terribly managed and are never efficient.) And then some figures for asphalt and cement in the US, as those are figures we know would be accurate.
I think most of the places we saw them they were spelled pretty similar to "pharmacy". It's a shame I didn't take a picture of the text so I would know for sure.
I thought we were talking about heavy rails, not light rails. Is the article talking about light rails?
Still, the article is interesting. The cost per mile is significantly higher for light rails than freeways, which is what I thought. Unfortunately it gives most of it's monetary figures based on theoretical and best case scenarios passengers per hour rather than typical usage. And even the numbers seem a little strange. The cited study says the showcase Calgary manages a cost of $2400 per weekday passenger (versus the more typical $12-30k) but that the cost is a mere $0.23 per ride. I never saw a ride anywhere near that cheap, even in Europe where they have passenger trains down to a science.
Still, it strongly illustrates the point that the viability of these systems is about moving the maximum number of people at once between two points, ala a freight train. Unfortunately most of the time you are nowhere near capacity, which completely throws off any idealistic cost quotes.
Any citations for power useage of trains versus other systems?
I see how you misunderstood what I said. Give the person a $20,000 car. Then use $14,000 to pay for all vehicle gas and maintenance. This would still be $1,000/year cheaper than the quoted $35,000/year, and far more useful.
This shouldn't be considered practical, which demonstrates just how impractical the situation InsaneProcessor was describing.
Railways are cheaper to construct than roads.
Do you have a citation for this? I thought it was the other way around, and some quick Googling didn't pull anything up for me.
That's not a fair comparison. You're going through one of the most expensive construction projects ever built (20 mile undersea tunnel), and you can't go by road anyway.
But you can go by plane, which is further, which I mentioned. You can also take a ferry, which is cheaper, and was also recommended to me by some european friends. But because of our schedule neither was practical, so we paid more.
Did you spend any time in any of the cities you visited? You can fit 500+ people on one train arriving in London at 8.30am. Would 500 more cars be better?
That's a good question. If they are coming from main Europe, planes would be more flexible and cheaper. There are a lot of factors to be considered, and I just don't know them all. I just know that trains aren't very flexible in where they go to/from, cost a lot energy to move quickly from one point to another, and seem to cost a fair amount of money.
I'm really hoping for a large scale personal rapid transit system to be more viable, but I doubt it will ever happen.
I would put the comfort of sleeping in my reclined car seat to be about the same or a little higher than one of the economic "6 people to a cabin" beds. And I can sit up in my car.
But it doesn't matter because there are specific reasons for taking a night train with a bed that driving a car can't compare with. I was just noting that I didn't find the sleeping platforms much of a "proper bed".
AFAIK, cost of building a mile of rail is far more expensive than a mile of asphalt, though there is less maintenance for the rail. (Life cost of concrete roads is much lower, but initial investment is high.) Rail is also not nearly as flexible in routing as roads. Those 4+ lage highways cost a lot, but rail systems are a completely different set of functionality.
I'd love to see the development of an extensive thin rail system using 3 passenger cars that are individually routeable. I think it'd be a happy median between the two.
Yes, but at $35k/year, it would be cheaper to buy all of the riders a $20,000 car and pay for gas, and get them a new car every year.
IF they decide this - as most municipalities in Europe have - then the answer to your question is "eh, no big deal, I can take the bus to the gym/grocery store."
I just got back from Europe and was surprised at the sheer number of "pharmacies" (marked with a green plus neon sign). These were easy to see, but if you looked closely were about as common as little food markets. Standing in front of my hotel in Paris could see three on the same street block as me.
People in Europe don't need to take the bus to the grocery store, and they don't buy groceries a week or so at a time. They have food markets a 100 feet from where they live, and they go there daily. Basic food distribution and practices are completely different in Europe than the US.
I did the same trip this past December. Prices were high, beds were terrible, with 6 beds in a room, dirty bathroom down at the car, and a German conductor that thought if she just talked louder in German we might understand the complex instructions she was giving us. And standing around in the unheated Paris train station for an hour in 20F temperatures was unpleasant.
But that's probably just the difference between economy and first class night train.
I think we must have different opinions on what constitutes a "proper bed".
Cheap asphalt with far more flexibility in where to go.
I'm from central Texas, and I recently spent a few weeks traveling by train in Europe. I would say that most of the train rides were far more expensive than making the same trip by car or plane. For instance, taking the Chunnel from London to Belgium was hundreds of dollars. But flying from Glasgow on Ryan Air would have been half the cost. The problem was that it wasn't really convenient to get to/from the airports. And renting a car wasn't practical because international car travel was expensive, especially when most of your destinations aren't very car friendly.
Big trains are great for slowly moving lots of heavy freight around, but it seems a bit lacking for moving lots of people around quickly.
Funny you should mention Adobe, we actually use Adobe Creative Suite Design Premium at work on a Terminal Server. We use a volume license from Adobe that has provisions specifically for use on terminal servers.
I would rank the experience of using Photoshop on it as mediocre. Everything works fine, except that there is a certain amount of latency that makes it annoying. Users using InDesign have never complained.
There were cost savings with this method. The licenses are more expensive, but we only needed 5 licenses for about 50 people as they generally aren't using the applications constantly. If we were purchasing for individual machines, we would have purchased only the applications they use instead of the entire suite, but this provides all applications to all users in the organization at all times. However if you have a graphic design artist that spends all day every day in Photoshop, it would make much more sense to just install it locally.
* We are actually using Citrix on our Terminal Servers, which I think is the source of the latency issues. Citrix is wonderful for people accessing their applications from home or while driving in a car, but it's use internally on a Gigabit LAN seems to add too much latency and unnecessary compression.
Nope. How many? Many companies offer licensing specifically for Terminal Services environments.
I think (not certain) that the reason you should prefix GNU to Linux is that Linux typically includes the entire GNU toolset to form an OS. The Linux kernel by itself is just a kernel. The Linux kernel with the GNU toolset is an actual OS, which is what most people are talking about. A program compiled with GCC though is just a program compiled with GCC.
I suspect they may have been doing things 'wrong', ie the "everything in one giant database" approach. It is impossible to tune a database to deal with all sizes of data efficiently, so the easiest thing to do is break the bits apart.
Have a database that stores the various important parts of the email header viewed by users. Extract the text from the emails and attachments and store it in a different database for performing text searches. Finally, store the full email header and email body in two different databases (or just the file system), indexed by a hash on their contents (this gets you single instance storage for when an email is sent to several people).
Doing this you can use different database systems on different servers that are carefully tuned to their specific workloads.
I recently had to build an online archive to display years worth of emails stored in PST files. I couldn't find anything simple so I just imported the emails into an IMAP account, and then wrote some PHP that queried and broke up the messages into pieces for storage. Granted, it's not serving gigabytes of emails to thousands of users simultaneously, but it's fast. It really makes me wonder how people get this wrong.
*I broke up the emails into their individual MIME attachments and stored those individually after compressing them with GZIP. This meant that broken emails might end up with missing pieces, but in my testing email clients never showed the broken pieces anyway. Also, most people just want the important parts of the header and the text, so the various MIME attachments are rarely retrieved so compressing them made a lot of sense.
Well, that's what I meant to imply. You're paying enough money that you don't debug the product, the vendor does. It's that way with most products at very large corporations.
I'm not sure I follow what you're talking about. What kind of "app" would one build on the Exchange platform? As far as I knew, it was a database for storing emails, contacts, and calendar information.
A properly set up Barracuda spam filter will let users log in and see any emails to their address that have been blocked. You can then deliver the mail, and set up whitelisting options.
I would rank our experience with Barracuda as "excellent". Your mileage may vary though as blocking conditions can depend a lot on how the administrator sets it up.
If you have over half a million users, Microsoft is going to have a team ready to fly out if there is a problem.
Database replication happens on write, not read.