Concede: the sample with a positive/negative stance have fallen as you say.
Reject: Out of 14000 papers, 99.7% climatologists agree Global Warming is man-made. This is the often-cited statistic: some 14000 papers, 99.7% agree on global warming.
My problem is that some 30% or 37% or so have agreed that Global Warming is real and man made; a fraction have rejected this outright; there's even one or two that say that global warming isn't a real thing; but then there's this huge majority that say there's not enough data to draw strong conclusions, that it appears that man-made climate change is probably a minor effect but could be negligible to non-existent or possibly bigger than they can conclude from the data. They're not saying that global warming and climate change aren't real, and they're not even saying it's not man-made; they're just saying that they've looked at the data and the data simply does not support a strong conclusion.
That's important: 31 scientists say the sky is falling and it's falling because we drink Coca-Cola. 68 scientists look up and say, "It looks like the clouds may be a little lower today... but clouds float up and down all the time. The measurements we've taken have shown some interesting deviations, and the manufacture and shipment of Coca-Cola seems to have some interesting potential to influence this, as well as some hit-and-miss correlation that we can't seem to pin down a model for. It seems possible, but uncertain at best, potentially unlikely, that Coca-Cola has a major effect on this lowering of cloud cover; we see other factors that at the moment look much more important. We can't rule out that the full effect of Coca-Cola on the sky falling is not yet understood, however--there is good evidence that we don't quite fully grasp it yet, since our models don't fucking work." You cannot then claim that science has "Concluded" that Coca-Cola is causing the sky to fall.
What's happened here is a minority opinion has been repeated loudly so many times that people think it's a scientific consensus; and the majority opinion is that the minority isn't necessarily wrong, just that they may be jumping the gun a little, as the evidence is non-conclusive. A lot of really smart people don't think this is a real, solid, scientific conclusion. In fact, most of them don't. They don't think it's quackery or wrong; they just don't have the scientific data to support this and reject other hypothesis, and so a lot of other hypothesis seem just as likely if not more likely, and so the answer is "We're not quite sure what's causing this, but we're more certain that this is a thing that's happening."
You don't strike me as unknowledgeable, so I'm going to go with A, but considering that the entire purpose of CDNs is to improve performance by decreasing traffic
TO THE ORIGIN.
If you have a 100Mbit/s link and a 20Mbit/s video stream, you cannot handle 1000 users at 20Mbit/s. So you find some people with gigabit links, you stick servers out there, you stream your origin to an edge origin at 20Mbit/s, which then streams out to the other edges at 20Mbit/s, which each can stream a gigabit to clients.
Do you know how much traffic there is going over the internet with 1000 users at 20Mbit/s? TWENTY GIGABITS. You don't decrease the traffic to 20Mbit. YOU send 20Mbit. THE INTERNET has to somehow carry 20Gbit.
If you stick a CDN node in the ISP--like Netflix does--then you're incurring some of that traffic over the ISP's last mile. Say 2Gbit out of that data center. Guess what? That ISP has to have capacity for 2Gbit out to their customers. That means if a customer downstream line has 100Mbit that's shared amongst 10 customers each throttled to 50Mbit, each pulling 20Mbit/, you have 200Mbit trying to come down a 100Mbit line. To make matters worse, you can't stick a CDN node in every ISP data center for every service; there's a lot of stuff just coming down the line, and we don't have any open standards to handle it all--much less all the live P2P stuff like GotoMeeting and Skype, which is a lesser concern.
This worked when it was Web browsing and 2 minute YouTube videos, the occasional P2P or CD download, etc. It worked for iTunes. The sudden uprising of hours on end of constant near-saturation, however, has made this model unviable and expensive.
If you put an antenna in your city, you can broadcast 10 miles. If you put an antenna 10 miles out that picks up that broadcast and re-broadcasts, then 10 miles away you have coverage there as well. Likewise, in that area 10 miles over there, the band you're broadcasting on is no longer usable FOR OTHER THINGS. If you blanket the earth with repeaters (this works for digital signals, less well for analog) and rebroadcast towers (i.e. digitally encode the signal or just ship it over cable, which has less degradation than radio), then i.e. Tokyo cannot have NHK9 because Chicago News 9 is taking up Channel 9.
People are divided over whether the last mile or the larger internet routes are more expensive to ISPs. They keep carrying this out to the extreme that one or the other doesn't matter, and the other is cripplingly expensive. Here's a hint: It's both. If you start sucking up last mile traffic, then the entire last mile model breaks and you need to severely beef up the last mile. If you start sucking up WAN line traffic, the same shit happens: infrastructure needs beefing, and you'll probably get less favorable terms on your next contract negotiation, and have to start charging your downstream users more. And you can't just stick a CDN at the last mile, on the pole, for everyone--technologically it's doable, but commercially it's expensive for SOMEONE and it's an implementation nightmare that's costly as hell.
Maybe we can do it one day. We'll need some new standards, new expectations, we'll need to get rid of all the fucking DRM sky is falling bullshit so that ISPs can do more off-the-shelf caching without Universal Studios crying that the ISPs are pirating their videos off NetFlix somehow. But we may find that the last mile is just going to get extremely fucking expensive and can't handle this. We may find that we're carrying too much over the backend when EVERYONE gets into the game. We may find a lot of things, but most of them we should expect now.
Beyond Eggs, however, are made from ingredients that include peas, sunflower lecithin, canola, and natural gums. They’re also gluten-free and cholesterol-free. In fact, the current formula is purportedly healthier than actual eggs.
Underestimating the health value of eggs. Assuming cholesterol is bad is hilariously stupid. The yolk of the egg is the healthiest part.
The blame should be lain on Tenable Network Security for being alarmist dumbasses.
In order to use HNAP, you have to be able to connect to the router. In other words, you need to be on the physical side of the network that supplies HNAP. This should not be WAN--it's WIFI or LAN. Wifi also has WPS.
WPS provides a short duration attack window for anyone within range to connect to your router. Short duration, user-initiated. It's your fault for pushing the button.
HNAP on the other hand requires someone to be on the network. Without HNAP, you could achieve the same by ARP flooding, ping scans, sniffing the network for packets (especially wifi), and so on. You'd find where packets are going, take a guess at the subnet mask (hint: it's/24), and so on.
Here's the punch line: HNAP is for homes, so the uncomfortable stuff above (ARP floods, ping scans! On a switched network you will not get far! Ha-HA!) is irrelevant. If I'm subject to your Ethernet switch preventing me from just sniffing packets and finding all the information, I'm probably inside your house OH SHIT!
And as for making it easier to probe for vulnerabilities and authentication bypass? I can just spoof packets from odd IPs and MAC addresses--if the router even bothers to block attack attempts--and throw every vulnerability at it in a few seconds. Find one that works, hook up to it.
Revealing information about a router so you can start probing for a vulnerability? Man, you can buy these off the shelves. People look for vulnerabilities all the time; then they note them down, find someone with the same software, and repeat the vulnerability. They also publish them online. They're not linking up to your wifi router and impromptu finding a previously undiscovered vulnerability in 20 minutes, man. Plus it's easier when you can actually log in: you log in, you read the network traffic looking at HTTP cookies and POST requests and such, and then you log out and see if you can A) magic up an authentication token; or B) do like on Clear's wireless routers and just POST commands un-authenticated, to which it responds by executing the command and telling you you're not logged in.
This is a pile of non-issue. Oh it's information-leaky alright; but the practical security implications are laughably moot.
A lot of things are hard to define. This is mostly because a strong definition of "fraud" and "bad faith" would be easy to compensate for and fraudulently abuse. "Misuse" and "Misappropriation" are strongly defined, but it is not wrong for an employee to misappropriate work resources for something he thinks is job related or an acceptable use by policy of those resources; it may carry consequences--more often no consequences in minor and isolated cases, if any even in extreme but still somehow reasonable cases--but it's called an "honest mistake". Intentional misuse and misappropriation of company resources is "employee fraud" and carries the heavy consequences of "fraud". Environments which do not distinguish between the two are highly stressful and often become exceedingly non-productive because people will avoid doing anything that doesn't have full 100% guarantee of being "allowed" and "correct".
"I'll know it when I see it" needs to be used sparingly, but it is important.
It's a 99.something% consensus, which is as solid as any consensus among a large population is ever going to get. Out of 13,950 peer-reviewed climate articles from 1991-2012, only 24 reject global warming. (source [desmogblog.com])
My god will you people stop this stupid shit? Out of 13,950 peer-reviewed climate articles from 1991-2012, only about 4000 concluded support for climate change. The rest claimed no strong evidence either way.
Look! Look at America! Out of 300 MILLION people, only like a few thousand complain about police state! You know what that means? By your metric it means 99.something% of America actively embraces the validity of a police state and has strong desire to implement one!
He said: "The climate will change. We must adapt."
You said: "Then we must stop changing the climate."
The problem with your statement is he included all the natural factors: The climate will eventually change outside of human influence, and humans will either need to adapt climate control technology or new behavioral patterns. You built in the assumption that humans are causing climate change and can stop climate change by simply stopping the human influence.
The problem is climate is changing; the earth isn't simply getting hotter. We're seeing areas that were cold become warm, areas which were warm becoming cold, and some areas having greater (hotter summers and colder winters) or lesser (cooler summers and hotter winters) extremes. As well, the seasons are moving about: some plants need 100 days or 50 days or such of freeze, and either a warm winter or a winter that's now balls-cold 40 days long instead of balls-cold 115 days long will cause them to die or stop bearing fruit.
With changes like this, humans don't simply find scarcity; rather they find change. A certain type of keystone crop can be grown more abundantly now, but humanity starves: the 280,000 acres of rice we were growing fail, and it previously would have been commercially unprofitable to take advantage of the new 350,000 acres of suitable land due to it being A) a little harder (more expensive to farm); and B) overproduction beyond the 280,000 acres worth of rice in the market, thus impossible to sell, especially at the inflated price from the cost of farming this. Likewise, maybe suddenly certain fruits and vegetables become more difficult to farm, but others become readily available--but we then need to take on the tremendous commercial undertaking of getting people to accept a new diet of strange and wonderful foods.
A large part of the problem is actually economic. That isn't to say that businesses are dumb and could just fix this by being not dumb; economic problems are real, and taking advantage of abundant resources while they're abundant is a good thing. To be sure, it's the best way to handle flow resources like solar energy: there's as much solar energy as there is, and it's going to go away whether you use it or not. Stock resources... there's something to be said about scarcity planning for things that will become scarce.
The difference between a stock resource and a flow resource is the time required for generation--energy can be produced in time, so anything that can be generated with energy also is a time problem (i.e. gasoline and oil are infinite resources: use solar energy to produce gasoline from atmospheric carbon and water vapor). If we're consuming the resource faster than the flow AND the resource can be directly accumulated (i.e. like oil, not like sunlight), it's a stock resource. If the resource cannot be directly accumulated or we're consuming it slower than we can accumulate it, it's a flow resource. Water for example is considered a flow resource, even though you can stock it: there is an infinite amount of water coming out of a given spring, but you only get so many liters per day. Water stored in barrels or in a basin (pond) is a stock resource.
Taking huge advantage of massive stock resources is economically wealth-generating, but you need a long term conservation plan. For example, oil. Back then, we didn't have the knowledge to do this; but if we had built massive solar generators using oil power and used them to generate oil from atmosphere, we'd be leveraging an OBSCENE amount of oil, an unconscionable amount really. But then we'd have these massive solar processing plants generating new oil, which becomes a flow resource; and we could store the oil in barrels and generate a stock. We can do that now, as well, which is quite likely a good alternative or supplement to electric cars--that is, energy-dense fuel cells or diesel generators as secondary power sources, fed from solar diesel.
If your power plant control systems fail, major cataclysmic failures occur.
If your GotoMeeting session disconnects, you look around and go, "... wut?" Then you look embarrassed at your press conference. Then you start it up again, and 45 seconds later you continue.
"Soft realtime" means it operates in real-time, able to land exact guarantees within given time constraints reliably; "Hard realtime" means it's able to make 100% perfect guarantees within given time constraints, and absolutely will not miss those guarantees. "Soft realtime" requirement means your shit only operates if realtime works, but it's okay if realtime fails--it can recover, or if it flat out fails out it can be restarted with no harm. "Hard realtime" requirements mean a failure is CATASTROPHIC and has SEVERE CONSEQUENCES.
Defining the requirements is a matter of risk management: a nuclear power plant's primary control systems are not hard realtime because they can survive a 45 minute failure interval; however the emergency recovery procedures are hard realtime because a failure has to be corrected within that interval. This is because the hard realtime requirement was too much of a risk at such short intervals, and so the reactors were designed to tolerate failure for a longer period than normal operation demands, and so can tolerate intermittent operational realtime failure and recover from it--even automatically recover--and otherwise can alert the human to DO SOMETHING ABOUT IT RIGHT NOW.
As you said, the purpose of a repeater is to help the signal go further, but it in no way reduces the amount of traffic over the network/airwaves (if anything, it increases it by increasing the range of the signal).
Yes, this puts more noise in the air. Mainly away from your transmission antenna, and non-interfering.
In contrast, CDNs lead to a marked decrease in traffic by caching the content locally so that it does not need to be pulled from the originating server again.
No, CDNs increase traffic. They decrease the number of clients connected to your main broadcast node. In the same way, a repeater or signal booster or a rebroadcast station (wired up with a cable) has people in that city over there listening to a signal from that antenna, rather than your main antenna. There is, however, more traffic over the whole fabric of the Internet--just as there is more radio traffic over the whole fabric of the airwaves with repeaters.
If you want to serve the geographical area of the city of New York from a single data center streaming video, you need like.. an OC-12. Maybe.
If you want to serve the geographical area of the fucking United States of America streaming video, you need an OC-192 or some kind of distributed streaming network (CDN).
Do you want to know how our live news streams go out? We have this shit software, it's called Wowza, it costs a thousand bucks and has nice features that are implemented terribly. It's a piece of shit but it fits our business case, so we spent $15,000 on licenses. Well now we have two Wowza live edge-origins nearby, and one Wowza live origin here. Live streaming news comes into our home network, where it is unicasted to the Wowza edge-origins. The edge-origins pick up the signal and re-unicast it to the other 9 servers all around the country, which are load balanced across using geographically weighted algorithms.
A radio repeater, you put an antenna up, and it picks up the radio signal from where you broadcast. It then boosts it, filters it if necessary/possible, and re-broadcasts.
These do the exact same thing. The fact that your attenuation on a radio wave is "distance and shit in the way" (shit in the way includes antennas, by the way: every antenna along the way has a minute effect on signal propagation--it weakens it), the attenuation on cable is "distance and shit picking up the signal, with quite a bit more effect from receivers actually attenuating part of the signal but still much of the effect having to do with raw distance", and the attenuation on Unicast network being exactly "shit picking up the signal, with very little but real and important effect via network distance--now measured by 'routers' in the way rather than linear transmission distance" doesn't change what we're looking at.
Your signal can service so many devices over so much distance. The longer the distance, the fewer devices--long distance on the Internet means little deviations in your 87 hops from Japan to Wisconsin add up to "only like 40 people can connect from Okinawa at once before they all start to have problems" while 400 people in Wisconsin start watching your live broadcast fine. If 500 people in Wisconsin max out the server, those 40 people in Japan suddenly can't get shit; well, if you can just barely get WXXX at 19 miles away, and then 40,000 houses in the first 4 miles put up YAGI antennas on the roof and start watching WXXX, suddenly your signal goes--that's a real thing, it happens, you lose signal because people closer pick it up first.
Yes bitch, they work the same fucking way. The coefficients change: impact of distance and impact of number of devices are different. It might be 0.1c + 0.9d = capacity for antenna broadcast, 0.25c + 0.75d = capacity for cable, and 0.95c + 0.05d = capacity for unicast IP (and distance is measured differently: by number of intervening routers rather than linear distance), but the basic model is the same: your
Well, yes. The correct response is for Comcast to continue to sell unlimited high-speed service that should function according to their models.
It's operational risk: Comcast assesses that they cannot be considered "dishonest" if they knowingly sell you an internet connection advertised as 50Mbit/s that actually functions as 50Mbit/s most of the time--almost all of the time--in real practice. They know that this satisfies the users, satisfies the lawyers, and satisfies their models of acceptable cost and revenue. If they over-saturate, users become unsatisfied; further, there come accusations of fraud, which lead to judicial investigation: it's one thing to over-provision in the honest belief that most of the time service comes as expected and when it doesn't it's very much brief and of no consequence to the user; it's quite different to over-provision and sell something that the user won't actually get half the time. "Bad faith".
The problem now is that everyone has become accustomed to high-speed internet, and has broken natural growth. Instead of medialets, images, sound, 2MB videos coming at people, and so on--an uptick in growth that should have come over half a decade and been, frankly, small--we've run into "I want to watch TV at 4000Kbit/s in 5 places at my house at once". Now everyone wants to be able to do that, nobody wants to pay $50/mo for 1Mbit service, nobody wants to pay $50/mo for 50Mbit service that disallows "heavy usage", and so on: something new has appeared, and now people believe they're entitled to it.
To avoid this, we could have just A) charged $750 or $1000 or $15,000 or so from the start for these $70 high-speed connections we enjoy today; B) never offered anyone anything faster than 128k ISDN; or C) provided usage caps, AOL "Internet Hours Per Month", or some other mechanism of telling people they're too retarded to drive the Internet Autobahn from the start. Which of these do you prefer?
Let's face it: Society has derived a lot of wealth from making full utilization of a resource that has been available up until now. What happened is exactly what happened with cars: We had all this oil and like 2000 cars in the world, so gasoline was cheap. Now there's 80 million cars, we're running out of oil, gasoline is expensive, people are either bullshitting about or trying to alert everyone to Global warming, politicians have muddied the environmental and economic issues so much that you can't tell if global warming or high oil prices are even real, and in general life is going to suck if we don't create an oil quota or make gasoline $8/liter or use wind/OTEC/solar-orbital to generate (expensive) gasoline.
Well the same happened with bandwidth: plenty of bandwidth to go around, a maximum amount of shit that can go on it at once but people only had demand for 1/1000 of that much shit for over a decade, so we just put 1000 people on the same chunk of bandwidth and it was good. Now people want to run 1000 times more shit down the line, and they wonder why it's backed up like a 4 inch pipe hooked up to the industrial dryer vent output of a laundromat.
I'm scratching my head at this because I think it requires some additional domain knowledge I don't have; but the overall set supplied seems to imply that your method of deduction is "inference". You don't have a solid argument; you have a scattered set of data that seems to show a big blank spot. This is how black holes are detected: something strange is happening here, so we infer that specific physical conditions exist that would cause that--we don't exactly know that's a real thing, but we think it is.
Right, we called those "Repeaters" and "Signal Boosters" in broadcast, since radio can go only so far and every receiver on a cable line attenuates the signal some.
The problem is it's not just Netflix. It's Netflix, YouTube, Hulu, Verizon Streaming TV, ePornoTube, RedStream Monthly, GayCentralUnlimited, AnimalPornHandjobIncorporated, Skype, Spotify, GotoMeeting, IronMountain Digital VPN Backup Services, iTunesVideo, Amazon Prime...
The Internet was designed for hypertext transport. Bulk things like images were tacked on. Even then, the high speed system that was built was built for rapid delivery of existing media with existing browsing habits.
Then: someone decided to bolt on streaming HD movies.
You ever see somebody bolt a side car onto a motorcycle? Well this is more like bolting a battleship to a Go Kart and expecting it to drag the ship down the street.
Now I'll explain in extended terms why you're retarded.
Let's assume there is enough bandwidth for 100Mbit/s.
Now let's assume you have 10 users. These users access in patterns of 10-15 second bursts in patterns with variance both at start and along the pattern--you may start browsing at a random time within 15 minutes of a known center, staying on pages for between 3 and 35 seconds, with occasional visits up to 3-5 minutes. Overall, the access is fairly random, and it's sets of 10-15 seconds of download.
Two choices.
Choice A: All 10 users get 100Mbit/s high-speed Internet. Occasionally, users will overlap and wind up with 3 users getting 33Mbit/s for like 30-45 seconds. Whoops. We can probably add 100 more users on this line with these access patterns, and nobody will much notice.
Choice B: We throttle every user to 10Mbit/s. They all have guaranteed, non-stop 10Mbit/s. If we add 10 more users, we need 100Mbit/s more bandwidth or we just sell 5Mbit/s service.
Choice A is superior for the end user: it's cheaper and it provides the same level of service.
UNTIL.
Suddenly those 10-15 second bursts become 2 hours of pulling 50Mbit/s video. Not just the 20 minutes of Windows Update that drags out 2 hours on Patch Tuesday, but constant hammering of available bandwidth.
Well, now you need 1000 times as much infrastructure and administrative overhead to manage this. Your $75/mo Comcast high speed cable bill? Yeah, that's going to be $75,000/mo now. It's just that expensive. We needed more tubes. 1000 times more tubes. Who the hell is going to pay for all these tubes?
So according to you, the correct response is to give everyone FiOS limited to dial-up speeds. 56K FiOS. $40/mo.
You think the Windows IP stack sniffs packets and then breaks TCP/IP to act as a DNS cache? Windows only has a list of recent DNS lookups because Windows apps call DnsQuery() to go through the WMI DNS Provider service, which is equivalent to setting 127.0.0.1 in/etc/resolv.conf on Linux and running bind locally as a caching nameserver.
Well, more "use libbind" the same as all the bind utils. It's one library call versus another. The point is that applications can, but don't, override this; there's nothing in the application allowing the user to tell it to do so, but it's bluntly easy for an application to provide a "stealth mode" that queries the OS for its DNS settings, then calls a stock DNS resolver library. Hell, I'm pretty sure that's how it's always done on Linux, using res_query() from libresolv (hence/etc/resolv.conf).
Slashdot Mirror
Concede: the sample with a positive/negative stance have fallen as you say.
Reject: Out of 14000 papers, 99.7% climatologists agree Global Warming is man-made. This is the often-cited statistic: some 14000 papers, 99.7% agree on global warming.
My problem is that some 30% or 37% or so have agreed that Global Warming is real and man made; a fraction have rejected this outright; there's even one or two that say that global warming isn't a real thing; but then there's this huge majority that say there's not enough data to draw strong conclusions, that it appears that man-made climate change is probably a minor effect but could be negligible to non-existent or possibly bigger than they can conclude from the data. They're not saying that global warming and climate change aren't real, and they're not even saying it's not man-made; they're just saying that they've looked at the data and the data simply does not support a strong conclusion.
That's important: 31 scientists say the sky is falling and it's falling because we drink Coca-Cola. 68 scientists look up and say, "It looks like the clouds may be a little lower today... but clouds float up and down all the time. The measurements we've taken have shown some interesting deviations, and the manufacture and shipment of Coca-Cola seems to have some interesting potential to influence this, as well as some hit-and-miss correlation that we can't seem to pin down a model for. It seems possible, but uncertain at best, potentially unlikely, that Coca-Cola has a major effect on this lowering of cloud cover; we see other factors that at the moment look much more important. We can't rule out that the full effect of Coca-Cola on the sky falling is not yet understood, however--there is good evidence that we don't quite fully grasp it yet, since our models don't fucking work." You cannot then claim that science has "Concluded" that Coca-Cola is causing the sky to fall.
What's happened here is a minority opinion has been repeated loudly so many times that people think it's a scientific consensus; and the majority opinion is that the minority isn't necessarily wrong, just that they may be jumping the gun a little, as the evidence is non-conclusive. A lot of really smart people don't think this is a real, solid, scientific conclusion. In fact, most of them don't. They don't think it's quackery or wrong; they just don't have the scientific data to support this and reject other hypothesis, and so a lot of other hypothesis seem just as likely if not more likely, and so the answer is "We're not quite sure what's causing this, but we're more certain that this is a thing that's happening."
I Science Thee unto Oblivion.
You don't strike me as unknowledgeable, so I'm going to go with A, but considering that the entire purpose of CDNs is to improve performance by decreasing traffic
TO THE ORIGIN.
If you have a 100Mbit/s link and a 20Mbit/s video stream, you cannot handle 1000 users at 20Mbit/s. So you find some people with gigabit links, you stick servers out there, you stream your origin to an edge origin at 20Mbit/s, which then streams out to the other edges at 20Mbit/s, which each can stream a gigabit to clients.
Do you know how much traffic there is going over the internet with 1000 users at 20Mbit/s? TWENTY GIGABITS. You don't decrease the traffic to 20Mbit. YOU send 20Mbit. THE INTERNET has to somehow carry 20Gbit.
If you stick a CDN node in the ISP--like Netflix does--then you're incurring some of that traffic over the ISP's last mile. Say 2Gbit out of that data center. Guess what? That ISP has to have capacity for 2Gbit out to their customers. That means if a customer downstream line has 100Mbit that's shared amongst 10 customers each throttled to 50Mbit, each pulling 20Mbit/, you have 200Mbit trying to come down a 100Mbit line. To make matters worse, you can't stick a CDN node in every ISP data center for every service; there's a lot of stuff just coming down the line, and we don't have any open standards to handle it all--much less all the live P2P stuff like GotoMeeting and Skype, which is a lesser concern.
This worked when it was Web browsing and 2 minute YouTube videos, the occasional P2P or CD download, etc. It worked for iTunes. The sudden uprising of hours on end of constant near-saturation, however, has made this model unviable and expensive.
If you put an antenna in your city, you can broadcast 10 miles. If you put an antenna 10 miles out that picks up that broadcast and re-broadcasts, then 10 miles away you have coverage there as well. Likewise, in that area 10 miles over there, the band you're broadcasting on is no longer usable FOR OTHER THINGS. If you blanket the earth with repeaters (this works for digital signals, less well for analog) and rebroadcast towers (i.e. digitally encode the signal or just ship it over cable, which has less degradation than radio), then i.e. Tokyo cannot have NHK9 because Chicago News 9 is taking up Channel 9.
People are divided over whether the last mile or the larger internet routes are more expensive to ISPs. They keep carrying this out to the extreme that one or the other doesn't matter, and the other is cripplingly expensive. Here's a hint: It's both. If you start sucking up last mile traffic, then the entire last mile model breaks and you need to severely beef up the last mile. If you start sucking up WAN line traffic, the same shit happens: infrastructure needs beefing, and you'll probably get less favorable terms on your next contract negotiation, and have to start charging your downstream users more. And you can't just stick a CDN at the last mile, on the pole, for everyone--technologically it's doable, but commercially it's expensive for SOMEONE and it's an implementation nightmare that's costly as hell.
Maybe we can do it one day. We'll need some new standards, new expectations, we'll need to get rid of all the fucking DRM sky is falling bullshit so that ISPs can do more off-the-shelf caching without Universal Studios crying that the ISPs are pirating their videos off NetFlix somehow. But we may find that the last mile is just going to get extremely fucking expensive and can't handle this. We may find that we're carrying too much over the backend when EVERYONE gets into the game. We may find a lot of things, but most of them we should expect now.
Road Rash 3D on PSOne has been eating my time. What is FTL?
How is it not false? No justification here. Both of you are rapidly masturbating your basal ganglia.
Beyond Eggs, however, are made from ingredients that include peas, sunflower lecithin, canola, and natural gums. They’re also gluten-free and cholesterol-free. In fact, the current formula is purportedly healthier than actual eggs.
Underestimating the health value of eggs. Assuming cholesterol is bad is hilariously stupid. The yolk of the egg is the healthiest part.
The blame should be lain on Tenable Network Security for being alarmist dumbasses.
In order to use HNAP, you have to be able to connect to the router. In other words, you need to be on the physical side of the network that supplies HNAP. This should not be WAN--it's WIFI or LAN. Wifi also has WPS.
WPS provides a short duration attack window for anyone within range to connect to your router. Short duration, user-initiated. It's your fault for pushing the button.
HNAP on the other hand requires someone to be on the network. Without HNAP, you could achieve the same by ARP flooding, ping scans, sniffing the network for packets (especially wifi), and so on. You'd find where packets are going, take a guess at the subnet mask (hint: it's /24), and so on.
Here's the punch line: HNAP is for homes, so the uncomfortable stuff above (ARP floods, ping scans! On a switched network you will not get far! Ha-HA!) is irrelevant. If I'm subject to your Ethernet switch preventing me from just sniffing packets and finding all the information, I'm probably inside your house OH SHIT!
And as for making it easier to probe for vulnerabilities and authentication bypass? I can just spoof packets from odd IPs and MAC addresses--if the router even bothers to block attack attempts--and throw every vulnerability at it in a few seconds. Find one that works, hook up to it.
Revealing information about a router so you can start probing for a vulnerability? Man, you can buy these off the shelves. People look for vulnerabilities all the time; then they note them down, find someone with the same software, and repeat the vulnerability. They also publish them online. They're not linking up to your wifi router and impromptu finding a previously undiscovered vulnerability in 20 minutes, man. Plus it's easier when you can actually log in: you log in, you read the network traffic looking at HTTP cookies and POST requests and such, and then you log out and see if you can A) magic up an authentication token; or B) do like on Clear's wireless routers and just POST commands un-authenticated, to which it responds by executing the command and telling you you're not logged in.
This is a pile of non-issue. Oh it's information-leaky alright; but the practical security implications are laughably moot.
A lot of things are hard to define. This is mostly because a strong definition of "fraud" and "bad faith" would be easy to compensate for and fraudulently abuse. "Misuse" and "Misappropriation" are strongly defined, but it is not wrong for an employee to misappropriate work resources for something he thinks is job related or an acceptable use by policy of those resources; it may carry consequences--more often no consequences in minor and isolated cases, if any even in extreme but still somehow reasonable cases--but it's called an "honest mistake". Intentional misuse and misappropriation of company resources is "employee fraud" and carries the heavy consequences of "fraud". Environments which do not distinguish between the two are highly stressful and often become exceedingly non-productive because people will avoid doing anything that doesn't have full 100% guarantee of being "allowed" and "correct".
"I'll know it when I see it" needs to be used sparingly, but it is important.
It's a 99.something% consensus, which is as solid as any consensus among a large population is ever going to get. Out of 13,950 peer-reviewed climate articles from 1991-2012, only 24 reject global warming. (source [desmogblog.com])
My god will you people stop this stupid shit? Out of 13,950 peer-reviewed climate articles from 1991-2012, only about 4000 concluded support for climate change. The rest claimed no strong evidence either way.
Look! Look at America! Out of 300 MILLION people, only like a few thousand complain about police state! You know what that means? By your metric it means 99.something% of America actively embraces the validity of a police state and has strong desire to implement one!
No, you're wrong.
He said: "The climate will change. We must adapt."
You said: "Then we must stop changing the climate."
The problem with your statement is he included all the natural factors: The climate will eventually change outside of human influence, and humans will either need to adapt climate control technology or new behavioral patterns. You built in the assumption that humans are causing climate change and can stop climate change by simply stopping the human influence.
The problem is climate is changing; the earth isn't simply getting hotter. We're seeing areas that were cold become warm, areas which were warm becoming cold, and some areas having greater (hotter summers and colder winters) or lesser (cooler summers and hotter winters) extremes. As well, the seasons are moving about: some plants need 100 days or 50 days or such of freeze, and either a warm winter or a winter that's now balls-cold 40 days long instead of balls-cold 115 days long will cause them to die or stop bearing fruit.
With changes like this, humans don't simply find scarcity; rather they find change. A certain type of keystone crop can be grown more abundantly now, but humanity starves: the 280,000 acres of rice we were growing fail, and it previously would have been commercially unprofitable to take advantage of the new 350,000 acres of suitable land due to it being A) a little harder (more expensive to farm); and B) overproduction beyond the 280,000 acres worth of rice in the market, thus impossible to sell, especially at the inflated price from the cost of farming this. Likewise, maybe suddenly certain fruits and vegetables become more difficult to farm, but others become readily available--but we then need to take on the tremendous commercial undertaking of getting people to accept a new diet of strange and wonderful foods.
A large part of the problem is actually economic. That isn't to say that businesses are dumb and could just fix this by being not dumb; economic problems are real, and taking advantage of abundant resources while they're abundant is a good thing. To be sure, it's the best way to handle flow resources like solar energy: there's as much solar energy as there is, and it's going to go away whether you use it or not. Stock resources... there's something to be said about scarcity planning for things that will become scarce.
The difference between a stock resource and a flow resource is the time required for generation--energy can be produced in time, so anything that can be generated with energy also is a time problem (i.e. gasoline and oil are infinite resources: use solar energy to produce gasoline from atmospheric carbon and water vapor). If we're consuming the resource faster than the flow AND the resource can be directly accumulated (i.e. like oil, not like sunlight), it's a stock resource. If the resource cannot be directly accumulated or we're consuming it slower than we can accumulate it, it's a flow resource. Water for example is considered a flow resource, even though you can stock it: there is an infinite amount of water coming out of a given spring, but you only get so many liters per day. Water stored in barrels or in a basin (pond) is a stock resource.
Taking huge advantage of massive stock resources is economically wealth-generating, but you need a long term conservation plan. For example, oil. Back then, we didn't have the knowledge to do this; but if we had built massive solar generators using oil power and used them to generate oil from atmosphere, we'd be leveraging an OBSCENE amount of oil, an unconscionable amount really. But then we'd have these massive solar processing plants generating new oil, which becomes a flow resource; and we could store the oil in barrels and generate a stock. We can do that now, as well, which is quite likely a good alternative or supplement to electric cars--that is, energy-dense fuel cells or diesel generators as secondary power sources, fed from solar diesel.
Investment, yes. Operating costs? Remember when Cogent de-peered from Quest because Quest was eating more bandwidth than they wanted to pay for?
Operational risk includes things like fraud indictment from bad faith, as well as bad consumer reputation.
So you say we'd all be better off if nobody had ever sold bigger than 56k ISDN?
If your brakes fail, you can die.
If your power plant control systems fail, major cataclysmic failures occur.
If your GotoMeeting session disconnects, you look around and go, "... wut?" Then you look embarrassed at your press conference. Then you start it up again, and 45 seconds later you continue.
"Soft realtime" means it operates in real-time, able to land exact guarantees within given time constraints reliably; "Hard realtime" means it's able to make 100% perfect guarantees within given time constraints, and absolutely will not miss those guarantees. "Soft realtime" requirement means your shit only operates if realtime works, but it's okay if realtime fails--it can recover, or if it flat out fails out it can be restarted with no harm. "Hard realtime" requirements mean a failure is CATASTROPHIC and has SEVERE CONSEQUENCES.
Defining the requirements is a matter of risk management: a nuclear power plant's primary control systems are not hard realtime because they can survive a 45 minute failure interval; however the emergency recovery procedures are hard realtime because a failure has to be corrected within that interval. This is because the hard realtime requirement was too much of a risk at such short intervals, and so the reactors were designed to tolerate failure for a longer period than normal operation demands, and so can tolerate intermittent operational realtime failure and recover from it--even automatically recover--and otherwise can alert the human to DO SOMETHING ABOUT IT RIGHT NOW.
As you said, the purpose of a repeater is to help the signal go further, but it in no way reduces the amount of traffic over the network/airwaves (if anything, it increases it by increasing the range of the signal).
Yes, this puts more noise in the air. Mainly away from your transmission antenna, and non-interfering.
In contrast, CDNs lead to a marked decrease in traffic by caching the content locally so that it does not need to be pulled from the originating server again.
No, CDNs increase traffic. They decrease the number of clients connected to your main broadcast node. In the same way, a repeater or signal booster or a rebroadcast station (wired up with a cable) has people in that city over there listening to a signal from that antenna, rather than your main antenna. There is, however, more traffic over the whole fabric of the Internet--just as there is more radio traffic over the whole fabric of the airwaves with repeaters.
If you want to serve the geographical area of the city of New York from a single data center streaming video, you need like.. an OC-12. Maybe.
If you want to serve the geographical area of the fucking United States of America streaming video, you need an OC-192 or some kind of distributed streaming network (CDN).
Do you want to know how our live news streams go out? We have this shit software, it's called Wowza, it costs a thousand bucks and has nice features that are implemented terribly. It's a piece of shit but it fits our business case, so we spent $15,000 on licenses. Well now we have two Wowza live edge-origins nearby, and one Wowza live origin here. Live streaming news comes into our home network, where it is unicasted to the Wowza edge-origins. The edge-origins pick up the signal and re-unicast it to the other 9 servers all around the country, which are load balanced across using geographically weighted algorithms.
A radio repeater, you put an antenna up, and it picks up the radio signal from where you broadcast. It then boosts it, filters it if necessary/possible, and re-broadcasts.
These do the exact same thing. The fact that your attenuation on a radio wave is "distance and shit in the way" (shit in the way includes antennas, by the way: every antenna along the way has a minute effect on signal propagation--it weakens it), the attenuation on cable is "distance and shit picking up the signal, with quite a bit more effect from receivers actually attenuating part of the signal but still much of the effect having to do with raw distance", and the attenuation on Unicast network being exactly "shit picking up the signal, with very little but real and important effect via network distance--now measured by 'routers' in the way rather than linear transmission distance" doesn't change what we're looking at.
Your signal can service so many devices over so much distance. The longer the distance, the fewer devices--long distance on the Internet means little deviations in your 87 hops from Japan to Wisconsin add up to "only like 40 people can connect from Okinawa at once before they all start to have problems" while 400 people in Wisconsin start watching your live broadcast fine. If 500 people in Wisconsin max out the server, those 40 people in Japan suddenly can't get shit; well, if you can just barely get WXXX at 19 miles away, and then 40,000 houses in the first 4 miles put up YAGI antennas on the roof and start watching WXXX, suddenly your signal goes--that's a real thing, it happens, you lose signal because people closer pick it up first.
Yes bitch, they work the same fucking way. The coefficients change: impact of distance and impact of number of devices are different. It might be 0.1c + 0.9d = capacity for antenna broadcast, 0.25c + 0.75d = capacity for cable, and 0.95c + 0.05d = capacity for unicast IP (and distance is measured differently: by number of intervening routers rather than linear distance), but the basic model is the same: your
Well, yes. The correct response is for Comcast to continue to sell unlimited high-speed service that should function according to their models.
It's operational risk: Comcast assesses that they cannot be considered "dishonest" if they knowingly sell you an internet connection advertised as 50Mbit/s that actually functions as 50Mbit/s most of the time--almost all of the time--in real practice. They know that this satisfies the users, satisfies the lawyers, and satisfies their models of acceptable cost and revenue. If they over-saturate, users become unsatisfied; further, there come accusations of fraud, which lead to judicial investigation: it's one thing to over-provision in the honest belief that most of the time service comes as expected and when it doesn't it's very much brief and of no consequence to the user; it's quite different to over-provision and sell something that the user won't actually get half the time. "Bad faith".
The problem now is that everyone has become accustomed to high-speed internet, and has broken natural growth. Instead of medialets, images, sound, 2MB videos coming at people, and so on--an uptick in growth that should have come over half a decade and been, frankly, small--we've run into "I want to watch TV at 4000Kbit/s in 5 places at my house at once". Now everyone wants to be able to do that, nobody wants to pay $50/mo for 1Mbit service, nobody wants to pay $50/mo for 50Mbit service that disallows "heavy usage", and so on: something new has appeared, and now people believe they're entitled to it.
To avoid this, we could have just A) charged $750 or $1000 or $15,000 or so from the start for these $70 high-speed connections we enjoy today; B) never offered anyone anything faster than 128k ISDN; or C) provided usage caps, AOL "Internet Hours Per Month", or some other mechanism of telling people they're too retarded to drive the Internet Autobahn from the start. Which of these do you prefer?
Let's face it: Society has derived a lot of wealth from making full utilization of a resource that has been available up until now. What happened is exactly what happened with cars: We had all this oil and like 2000 cars in the world, so gasoline was cheap. Now there's 80 million cars, we're running out of oil, gasoline is expensive, people are either bullshitting about or trying to alert everyone to Global warming, politicians have muddied the environmental and economic issues so much that you can't tell if global warming or high oil prices are even real, and in general life is going to suck if we don't create an oil quota or make gasoline $8/liter or use wind/OTEC/solar-orbital to generate (expensive) gasoline.
Well the same happened with bandwidth: plenty of bandwidth to go around, a maximum amount of shit that can go on it at once but people only had demand for 1/1000 of that much shit for over a decade, so we just put 1000 people on the same chunk of bandwidth and it was good. Now people want to run 1000 times more shit down the line, and they wonder why it's backed up like a 4 inch pipe hooked up to the industrial dryer vent output of a laundromat.
I'm scratching my head at this because I think it requires some additional domain knowledge I don't have; but the overall set supplied seems to imply that your method of deduction is "inference". You don't have a solid argument; you have a scattered set of data that seems to show a big blank spot. This is how black holes are detected: something strange is happening here, so we infer that specific physical conditions exist that would cause that--we don't exactly know that's a real thing, but we think it is.
Right, we called those "Repeaters" and "Signal Boosters" in broadcast, since radio can go only so far and every receiver on a cable line attenuates the signal some.
The problem is it's not just Netflix. It's Netflix, YouTube, Hulu, Verizon Streaming TV, ePornoTube, RedStream Monthly, GayCentralUnlimited, AnimalPornHandjobIncorporated, Skype, Spotify, GotoMeeting, IronMountain Digital VPN Backup Services, iTunesVideo, Amazon Prime...
Video is a soft-reatime system, but it's realtime alright.
The Internet was designed for hypertext transport. Bulk things like images were tacked on. Even then, the high speed system that was built was built for rapid delivery of existing media with existing browsing habits.
Then: someone decided to bolt on streaming HD movies.
You ever see somebody bolt a side car onto a motorcycle? Well this is more like bolting a battleship to a Go Kart and expecting it to drag the ship down the street.
the biggest cost for ISPs is the last mile
Who told you that?
Now I'll explain in extended terms why you're retarded.
Let's assume there is enough bandwidth for 100Mbit/s.
Now let's assume you have 10 users. These users access in patterns of 10-15 second bursts in patterns with variance both at start and along the pattern--you may start browsing at a random time within 15 minutes of a known center, staying on pages for between 3 and 35 seconds, with occasional visits up to 3-5 minutes. Overall, the access is fairly random, and it's sets of 10-15 seconds of download.
Two choices.
Choice A: All 10 users get 100Mbit/s high-speed Internet. Occasionally, users will overlap and wind up with 3 users getting 33Mbit/s for like 30-45 seconds. Whoops. We can probably add 100 more users on this line with these access patterns, and nobody will much notice.
Choice B: We throttle every user to 10Mbit/s. They all have guaranteed, non-stop 10Mbit/s. If we add 10 more users, we need 100Mbit/s more bandwidth or we just sell 5Mbit/s service.
Choice A is superior for the end user: it's cheaper and it provides the same level of service.
UNTIL.
Suddenly those 10-15 second bursts become 2 hours of pulling 50Mbit/s video. Not just the 20 minutes of Windows Update that drags out 2 hours on Patch Tuesday, but constant hammering of available bandwidth.
Well, now you need 1000 times as much infrastructure and administrative overhead to manage this. Your $75/mo Comcast high speed cable bill? Yeah, that's going to be $75,000/mo now. It's just that expensive. We needed more tubes. 1000 times more tubes. Who the hell is going to pay for all these tubes?
So according to you, the correct response is to give everyone FiOS limited to dial-up speeds. 56K FiOS. $40/mo.
You think the Windows IP stack sniffs packets and then breaks TCP/IP to act as a DNS cache? Windows only has a list of recent DNS lookups because Windows apps call DnsQuery() to go through the WMI DNS Provider service, which is equivalent to setting 127.0.0.1 in /etc/resolv.conf on Linux and running bind locally as a caching nameserver.
Well, more "use libbind" the same as all the bind utils. It's one library call versus another. The point is that applications can, but don't, override this; there's nothing in the application allowing the user to tell it to do so, but it's bluntly easy for an application to provide a "stealth mode" that queries the OS for its DNS settings, then calls a stock DNS resolver library. Hell, I'm pretty sure that's how it's always done on Linux, using res_query() from libresolv (hence /etc/resolv.conf).
What blogging platform hosts thousands of blogs under a single domain name?
Applications can override that. Just connect to 8.8.8.8 instead of running OperatingSystemApiDNSLookup().