We happen to be in a building with 5 (last count) OC-192s for another tenant from multiple providers. (Video game testing/development or something.) Their incremental cost of service is pretty attractive, and we have a long term contract.
$60-70 for ATT DSL, $400 for 10Mbit fiber (and I think $600 for 100Mbit).
To me the value of the fiber wins hands-down; we get about 10x performance for 7x cost, or 100x performance for 10x cost of DSL. Service is more reliable, lower latency, symmetric, and managed. The symmetric starts to matter a lot for things like the Microcells, VOIP, and video conferencing.
At our office we have fiber from XO, TWTelecom, Abovenet, and a few other smaller players. Time Warner is spending ungodly sums to bring fiber down the corridor to serve ~5MM square feet of offices.
But, ATT only offers "up to 6mbit" DSL. Pricing is comparable for value, but the offering is simply not up to snuff.
The devices you mention have a 1:100 chance of needing to operate in a 25-year timeframe for a single instance, roughly. They each add less than 1% to the cost of doing business.
The TSA installation for a single checkpoint process on average 30,000 passengers per year, and has an annual operating cost of (roughly) $600,000, or say $40/round-trip ticket. That is closer to 10% average airfare. It's probability of detecting a "terrorist" (let's go ahead and take a liberal interpretation of the term to include anyone intending to do harm with a weapon on a plane that cannot be detected by a simple metal detector) is about 1:400,000,000,000. (The pre 9/11 checkpoints would have had an annual operating cost closer to $150,000 in fairness.). There is also significant economic cost due to the silly restrictions they create, but not sure how to capture that.
So, you have a substantial cost for no significant improvement in security delivered. Sounds like security theater to me.
Not really. The DJI Phantom-2 is famous for thinking the battery is dead and splashing straight down.
Waypoints make these things pretty cool, but you really need to keep the size down before they get "too big", and that is hard when you need better dampening on the gimbal, better low light performance, better endurance, and wider control radius.
I have a couple of the Asus routers, and I love them. One runs as an openvpn server, the other runs a few services to simplify remote administration of an offsite location. Good little boxes.
But, it has really opened my eyes as to how bad security can be. These systems are at least slightly more secure than the WD drives. Third party firmware adds some levels of complexity, but a whole lot of functionality.
Redundancy can be nice, but I agree that the best solution to the consolidation is municipal fiber. It won't be cost competitive for underground utilities, but it seems like the only solution at this point. Co-op's are nice, but they could still be bought out or compromised.
The US decided it is a commodity rather than a currency. That simplifies things pretty much and more accurately reflects what it is. Maybe a derivative of a commodity is more reasonable, but that is more of a detail.
Personally, I have a few gold coins, because the portability is nice. I'm not really worried about the inherent value of the coins when gold eventually comes back down to reality pricing, because it is still an easy way to be able to move money around. Much of the downside of bitcoins also relate pretty well to gold, only the redemption network is much more limited.
If/when Mt Gox crashes, Bitcoins will go back down to a more stable pricing. Eventually, expect the entrenched financial firms to get into the race, which will ultimately be the downfall. Bitcoin ETF... coming to a stock exchange near you...
A very low percentage of IT people understand security issues to a sufficient degree to be able to act on them in the abstract. Talking to the director of IT at a very large defense contractor a few years back about a new proposed SCADA network, I showed him the plan for our isolated network, and the proxy/firewall connection to his corporate network, and asked him how they wanted to treat it. He was prompt to ask who needed access, how much throughput would it need, and if we needed more than one IP address.
I then went into my laundry list of bigger issues, so he suggested they just get a dedicated DSL line for it so it didn't need to be connected to the corporate network, and just make the SCADA vendor responsible for security!
People want to put security issues into buckets. The problem is that issues today are substantially more complicated than that.
Just look at slashdot beta... That is what "news for nerds" is trying to be now. Lowest common denominator only, please.
Beta is not usable on my laptop. It is not useable on my 27" desktop. It is not useable on my tablet. It is not useable on my phone. The problems you see on mobile still happen on a 27" monitor with a fullscreen browser.
The images are what many of us here consider "dumbing down" the site. It is tricky, but things eventually turn into sites like sfgate.com, where all the news is suddenly a slideshow without any actual content. Being able to integrate photos into the summary makes some sense for a few of the stories. I remember trying to find an old story about the Corel Netwinder (wow... was that really 15 years ago?!) and wanted to see a picture of the thing. (Couldn't remember the name of the product at the time.) But, 99% of what people are reading are text comments, so how important is it really?
For an example of what some of this crap leads to, take a look at airliners.net. A very miserable experience in a discussion forum, which a few very simple things keep/. from becoming. A few of us in the "audience" do actually care about that stuff.
I'll offer a few of my thoughts for Timothy and Soulskill: -Content first. Information density. No video! -Quality and relevance gives value to content. Relevance in deep nested threads is still relevance; the existing system today does a great job there. -Information portability. Ease of access, be it via search engine of a different device... or read outloud by a robot overlord.
Good luck; seems like you have your work cut out for you!
No, the site needs to be more mobile-friendly as well.
I hated when the "classic" interface eliminated the hard-coded thresholds; I used to love checking in to a discussion well in progress with clear, insightful information promoted and the noise hidden away. Unfortunately, I do appreciate that this undermines the most important thing of/. : Community. Everything is always in flux. So, I am ok with it now. There are still things in the old-timer's interface I miss though.
By the way, the phone number for Dice Holding's corporate headquarters is 212-725-6550. Be sure to give them a call and let them know what you think about their new design.
No, it is that proper security is really hard to do, especially when you deal with third parties that need to access portions of the network that management also needs to access. It doesn't help when the third party has one company account, and a reasonably high turnover rate of employees.
I used to have a rolodex of access cards for different clients and sites. Many companies required a different card for each building. Then this magical internet came along and they merged all of the security systems into central corporate security. Like magic I only needed one card for each client, locked down to specific areas I needed access in different building. Then... they had a problem. I couldn't get into the building to help out. It wasn't the end of the world, but the project manager I was working for ended up giving me all access to keep it from happening again. It took two years for a corporate security audit to call me and ask why the hell I needed "ring zero access" or whatever they called it. Up until that I had cash vault access for whatever stupid reason.
The bigger and more distributed organizations get, and the deeper the tree is on the contractors they work with, the more it becomes impossible to manage security without paying a huge efficiency penalty.
Sorry to get so off-topic; aren't we supposed to be talking about how miserable the beta.slashdot.org site is? Completely unusable; are there any other competing websites that could resurrect the old slashcode?
It's a shame there is such a disconnect./. Really was the start of Web 2.0. Looks like it will be the start of 3.0ooohpleezewhybother as well.
I'm sticking/. into the hosts file next week to resist temptation. Pretty clear management doesn't read slashdot though. The web designers must have gotten their feelings hurt and stopped as well...
At work with 10-Meg metropolitan fiber, no problem. At home, I pay 10% the price for a 15/5 plan; I would be kidding myself if I thought I could get the same level of service. The $ savings justifies it though, within reason.
I think Verizon's business FIOS plans are about double residential rates, which should give you less contention.
If you are throttling an hour a day it isn't QoS anymore, it is traffic shaping. If you throttle an hour a day to levels below 10% of rated bandwidth, are you delivering the bandwidth promised?
It is one thing for an ISP to react to using 100% bandwidth 100% of the time, but another altogether when you want peak capacity for 5% of the time, 50% for 15%, and 10% for the remaining 80% of the time. If you run two Netflix streams at a time at home with a 2 Mbit connection, you have unreasonable expectations of your ISP. Running two with a 10 Mbit connection doesn't seem as far off, as long as you aren't talking 24x7.
Yes it can be done. No, it isn't efficient. If you are only paying $0.0075/kWh in fuel (and ignoring disposal costs), you can afford some inefficiencies, but the relative advantage of nuclear goes away pretty quickly if your efficiency drops off dramatically.
That is actually done to a very large extent now. Foundries running electric arc furnaces or induction furnaces only run in the off-peak period, currently at night. This artificially increases base loads.
The problem with trying to match generation with demand is that you still have a transmission/distribution problem. Distributed generation is the only way to really solve that, and again economics make it difficult to distribute power generation to the point where local demand is matched to local production in both capacity and timing.
People are trying to get closer to this-- automated demand response can help a little bit.
The California ISO is pretty open with information. They track daily anticipated demand, actual demand, and available capacity. Some actually predict that solar energy that is not time-shifted will become nearly worthless in five years.
Exactly what I was going to post; fundamentally variable production and nuclear don't mix well. Nuclear only works as a base load power source.
Presumably you could design a nuclear reactor and turbine system to be able to better modulate to match load. This wouldn't fix nuclear power's economics problems though, which are really what force it to be running all-out as many hours a year as they can.
Slashdot Mirror
We happen to be in a building with 5 (last count) OC-192s for another tenant from multiple providers. (Video game testing/development or something.) Their incremental cost of service is pretty attractive, and we have a long term contract.
$60-70 for ATT DSL, $400 for 10Mbit fiber (and I think $600 for 100Mbit).
To me the value of the fiber wins hands-down; we get about 10x performance for 7x cost, or 100x performance for 10x cost of DSL. Service is more reliable, lower latency, symmetric, and managed. The symmetric starts to matter a lot for things like the Microcells, VOIP, and video conferencing.
At our office we have fiber from XO, TWTelecom, Abovenet, and a few other smaller players. Time Warner is spending ungodly sums to bring fiber down the corridor to serve ~5MM square feet of offices.
But, ATT only offers "up to 6mbit" DSL. Pricing is comparable for value, but the offering is simply not up to snuff.
The devices you mention have a 1:100 chance of needing to operate in a 25-year timeframe for a single instance, roughly. They each add less than 1% to the cost of doing business.
The TSA installation for a single checkpoint process on average 30,000 passengers per year, and has an annual operating cost of (roughly) $600,000, or say $40/round-trip ticket. That is closer to 10% average airfare. It's probability of detecting a "terrorist" (let's go ahead and take a liberal interpretation of the term to include anyone intending to do harm with a weapon on a plane that cannot be detected by a simple metal detector) is about 1:400,000,000,000. (The pre 9/11 checkpoints would have had an annual operating cost closer to $150,000 in fairness.). There is also significant economic cost due to the silly restrictions they create, but not sure how to capture that.
So, you have a substantial cost for no significant improvement in security delivered. Sounds like security theater to me.
Not really. The DJI Phantom-2 is famous for thinking the battery is dead and splashing straight down.
Waypoints make these things pretty cool, but you really need to keep the size down before they get "too big", and that is hard when you need better dampening on the gimbal, better low light performance, better endurance, and wider control radius.
...oh the irony.
I have a couple of the Asus routers, and I love them. One runs as an openvpn server, the other runs a few services to simplify remote administration of an offsite location. Good little boxes.
But, it has really opened my eyes as to how bad security can be. These systems are at least slightly more secure than the WD drives. Third party firmware adds some levels of complexity, but a whole lot of functionality.
Redundancy can be nice, but I agree that the best solution to the consolidation is municipal fiber. It won't be cost competitive for underground utilities, but it seems like the only solution at this point. Co-op's are nice, but they could still be bought out or compromised.
The US decided it is a commodity rather than a currency. That simplifies things pretty much and more accurately reflects what it is. Maybe a derivative of a commodity is more reasonable, but that is more of a detail.
Personally, I have a few gold coins, because the portability is nice. I'm not really worried about the inherent value of the coins when gold eventually comes back down to reality pricing, because it is still an easy way to be able to move money around. Much of the downside of bitcoins also relate pretty well to gold, only the redemption network is much more limited.
If/when Mt Gox crashes, Bitcoins will go back down to a more stable pricing. Eventually, expect the entrenched financial firms to get into the race, which will ultimately be the downfall. Bitcoin ETF... coming to a stock exchange near you...
A very low percentage of IT people understand security issues to a sufficient degree to be able to act on them in the abstract. Talking to the director of IT at a very large defense contractor a few years back about a new proposed SCADA network, I showed him the plan for our isolated network, and the proxy/firewall connection to his corporate network, and asked him how they wanted to treat it. He was prompt to ask who needed access, how much throughput would it need, and if we needed more than one IP address.
I then went into my laundry list of bigger issues, so he suggested they just get a dedicated DSL line for it so it didn't need to be connected to the corporate network, and just make the SCADA vendor responsible for security!
People want to put security issues into buckets. The problem is that issues today are substantially more complicated than that.
Just look at slashdot beta... That is what "news for nerds" is trying to be now. Lowest common denominator only, please.
I'll give the contrarian viewpoint then.
Beta is not usable on my laptop. It is not useable on my 27" desktop. It is not useable on my tablet. It is not useable on my phone. The problems you see on mobile still happen on a 27" monitor with a fullscreen browser.
The images are what many of us here consider "dumbing down" the site. It is tricky, but things eventually turn into sites like sfgate.com, where all the news is suddenly a slideshow without any actual content. Being able to integrate photos into the summary makes some sense for a few of the stories. I remember trying to find an old story about the Corel Netwinder (wow... was that really 15 years ago?!) and wanted to see a picture of the thing. (Couldn't remember the name of the product at the time.) But, 99% of what people are reading are text comments, so how important is it really?
For an example of what some of this crap leads to, take a look at airliners.net. A very miserable experience in a discussion forum, which a few very simple things keep /. from becoming. A few of us in the "audience" do actually care about that stuff.
I'll offer a few of my thoughts for Timothy and Soulskill:
-Content first. Information density. No video!
-Quality and relevance gives value to content. Relevance in deep nested threads is still relevance; the existing system today does a great job there.
-Information portability. Ease of access, be it via search engine of a different device... or read outloud by a robot overlord.
Good luck; seems like you have your work cut out for you!
No, the site needs to be more mobile-friendly as well.
I hated when the "classic" interface eliminated the hard-coded thresholds; I used to love checking in to a discussion well in progress with clear, insightful information promoted and the noise hidden away. Unfortunately, I do appreciate that this undermines the most important thing of /. : Community. Everything is always in flux. So, I am ok with it now. There are still things in the old-timer's interface I miss though.
Hey, I am quite happy with my slashdot.org/palm on Lynx, thank you very much!
There is always the approach of calling Dice Holdings. Their telephone number is 212-725-6550.
Damn FDIC. Oh well...
By the way, the phone number for Dice Holding's corporate headquarters is 212-725-6550. Be sure to give them a call and let them know what you think about their new design.
Time for a firefox extension "classic slashdot?"
No, it is that proper security is really hard to do, especially when you deal with third parties that need to access portions of the network that management also needs to access. It doesn't help when the third party has one company account, and a reasonably high turnover rate of employees.
I used to have a rolodex of access cards for different clients and sites. Many companies required a different card for each building. Then this magical internet came along and they merged all of the security systems into central corporate security. Like magic I only needed one card for each client, locked down to specific areas I needed access in different building. Then... they had a problem. I couldn't get into the building to help out. It wasn't the end of the world, but the project manager I was working for ended up giving me all access to keep it from happening again. It took two years for a corporate security audit to call me and ask why the hell I needed "ring zero access" or whatever they called it. Up until that I had cash vault access for whatever stupid reason.
The bigger and more distributed organizations get, and the deeper the tree is on the contractors they work with, the more it becomes impossible to manage security without paying a huge efficiency penalty.
Sorry to get so off-topic; aren't we supposed to be talking about how miserable the beta.slashdot.org site is? Completely unusable; are there any other competing websites that could resurrect the old slashcode?
Is Kuro5hin still around? ;)
It's a shame there is such a disconnect. /. Really was the start of Web 2.0. Looks like it will be the start of 3.0ooohpleezewhybother as well.
I'm sticking /. into the hosts file next week to resist temptation. Pretty clear management doesn't read slashdot though. The web designers must have gotten their feelings hurt and stopped as well...
At work with 10-Meg metropolitan fiber, no problem. At home, I pay 10% the price for a 15/5 plan; I would be kidding myself if I thought I could get the same level of service. The $ savings justifies it though, within reason.
I think Verizon's business FIOS plans are about double residential rates, which should give you less contention.
So, can we have a user's buyout of /. then? I'll pledge $20 for first year operating costs.
Too bad Taco's new site is even worse than Beta...
If you are throttling an hour a day it isn't QoS anymore, it is traffic shaping. If you throttle an hour a day to levels below 10% of rated bandwidth, are you delivering the bandwidth promised?
It is one thing for an ISP to react to using 100% bandwidth 100% of the time, but another altogether when you want peak capacity for 5% of the time, 50% for 15%, and 10% for the remaining 80% of the time. If you run two Netflix streams at a time at home with a 2 Mbit connection, you have unreasonable expectations of your ISP. Running two with a 10 Mbit connection doesn't seem as far off, as long as you aren't talking 24x7.
Yes it can be done. No, it isn't efficient. If you are only paying $0.0075/kWh in fuel (and ignoring disposal costs), you can afford some inefficiencies, but the relative advantage of nuclear goes away pretty quickly if your efficiency drops off dramatically.
Don't tell that to Toyota! They are investing heavily in Hydrogen fuel cells for SUV-class cars.
That is actually done to a very large extent now. Foundries running electric arc furnaces or induction furnaces only run in the off-peak period, currently at night. This artificially increases base loads.
The problem with trying to match generation with demand is that you still have a transmission/distribution problem. Distributed generation is the only way to really solve that, and again economics make it difficult to distribute power generation to the point where local demand is matched to local production in both capacity and timing.
People are trying to get closer to this-- automated demand response can help a little bit.
The California ISO is pretty open with information. They track daily anticipated demand, actual demand, and available capacity. Some actually predict that solar energy that is not time-shifted will become nearly worthless in five years.
Exactly what I was going to post; fundamentally variable production and nuclear don't mix well. Nuclear only works as a base load power source.
Presumably you could design a nuclear reactor and turbine system to be able to better modulate to match load. This wouldn't fix nuclear power's economics problems though, which are really what force it to be running all-out as many hours a year as they can.