> Drop them at random and let the end user see the stupid design decision made by their company of choice.
You could do that, but Netflix or Youtube doesn't work well on your ISP, the customers don't yell at Youtube. And when the server retransmits the packets you dropped at random, it makes the network more congested for *everyone*.
"Just make it worse for everyone" doesn't sound like the best idea to me.
Zero rating does bring up a couple of important issues, agreed.
I think there has been confusion all around. A few years ago, the big controversy around "network neutrality" was basically that Netflx wanted to be the only web site in the world who didn't have to pay their hosting bill. They intentionally confused that with network neutrality and many people on Slashdot, perhaps most, fell for it. That's a *completely* separate issue from anything related to zero rating.
> your stance that carriers ought to be able to do anything they want with the traffic
I've said the exact opposite several times in this thread. I've said I think we need specific rules directed at specific issues.
> why don't you volunteer to write the legal text that would enable the objectives of net neutrality
(and everyone here, including you, knows damn well what those objectives are)
I note you used the plural objectiveS. That's insightful given that other people posting in this thread have referred to the OBJECTIVE, singular, but three different people posting have said three different objectives. If you want to address zero rating, make a rule about zero rating. If peering agreements are a problem, make a rule about that. If providing different services to paid hosting customers than you do for unpaid, unrelated companies is a problem, that's a separate rule.
The issue that's probably most clear to me is if Comcast was artificially slowing Netflix traffic at one point in time. Maybe probably they were - only certain locations experienced the issue, suggesting that the cause *may* have been related to peering, a legitimate business negotiation. It's a lot simpler to address *that* issue, if they were artificially throttling it within Comcast's network, without conflating it with zero rating and other issues.
The discussion of NN heated up, and started appearing on the Slashdot regularly, a few years ago when Netflix, which was smaller at the time, was basically demanding transit service for free. They thought they should be the only web site in the world who doesn't have to pay their web hosting bill. "Everybody here knows", but some here were suckered by Netflix into confusing that peering negotiation with network neutrality. On *that*, Netflix's wish for free peering, I don't think we need any regulation - they can negotiate those arrangements just like every other business in the world does.
A trickier issue is this. Suppose I have a server with some videos on it and I want to stream those to the internet. Obviously I need to connect that server to the internet, through some kind of ISP. So maybe I go to Comcast Business and buy a 100Mbps or gigabit connection for my server(s). That's standard procedure for how you set up a web site. I pay my monthly bill, they sell me a gigabit of connectivity to their network. In real life, I currently have servers in three locations, connected to three ISPs (the B2B side of the ISPs, selling dedicated bandwidth to their network and transit off of their network). Like anyone else hosting a web site of any significant size, there are two numbers that describe how much bandwidth I get, basically the minimum guaranteed and the maximum possible (CIR and signaling rate). I decide how much guaranteed bandwidth I want to buy and what signaling rate I want. The guaranteed bandwidth (CIR) can be guaranteed only because it's prioritized over any non-guaranteed bandwidth. Does that make sense so far? Thr point is that to host servers, you pay for two bandwidth rates - minimum guaranteed and maximum possible. Everybody's minimum bandwidth is prioritized over everybody's non-guaranteed portion. So...
Suppose Hulu does the exact same thing that everyone with a large web site does and buys an internet connection for one of their servers. They happen to buy this connection from Time Warner. Picking from the price list, they choose 1Gbps CIR (guaranteed) on a 10Gbps connection. That's a normal web hosting contract. But... fuck! It's also "paid prioritization"! Many people think "paid prioritization" should be illegal, but there's no technical difference between "paid prioritization" and "web hosting". I'm not sure *how* you'd write that law. Do you say that any given web site is only allowed to have servers at one location, so they aren't buying hosting (with traffic guarantees) from several different ISPs? That seems like a bad solution. There's literally no difference between me paying Time Warner to host my one server for raysvideos.com and Netflix paying them to host a cache server. Do *you* have any ideas for an effective rule on that topic?
Someone who has truly mastered their craft may perfection 99% of the time. Or not - Tom Brady completes 64% of his passes.
Suppose the Firefox programmers were the most competent human beings to ever walk the earth, and got it right 99.99% of the time. With 14 million lines of code, they would have 14,000 flaws.
On the other hand, if the Rust string handling functions don't permit buffer overflows, they don't permit buffer overflows - ever. You can't write a buffer overflow in a language that doesn't use buffers. Not only will there not *be* such errors, but you can *prove* there are no such errors, you can trust it.
I don't have any opinion on Rust specifically, good or bad. I'm sure it has tradeoffs. The idea that you shouldn't use reliable tools because humans should just be perfect os silly.
> In practice you can too. Net neutrality is about the source of the data. QoS is about the content. They are very easily distinguished
Okay so I've got some packets from 45.83.129.42. I can tell it's some kind of video. Maybe it's a live teleconference, meaning delay would be really bad and any late packets need to be dropped - they won't be used anyway. Or maybe it's a pre-recorded video and the client is caching 30 seconds, so delay doesn't matter and late packets should be delivered and even retried several times. *We don't know.* How should I treat those packets? When a packet is late should I drop it, forward it, retry it?
I also have some packets from Netflix (actually most of my packets are from Netflix. I *do* know it's not a live teleconference, and latency doesn't matter. I even know how the client-side caching works. How should I treat this flow? Should I pretend I'm stupid and drop late packets, causing the sender to retry them and slow down the network for everyone?
When you make shaping and routing decisions, you can trade bandwidth for packet loss and latency for jitter, on a flow-by-flow basis. I don't need to upgrade anything in order to deliver the right mix of jitter, rate, loss, and latency that works well for Netflix streams. The only reason I wouldn't be able to deliver what customers want is that some of you want to make it illegal for me to do so.
I understand your point, I believe I know what you want.
I think what your missing is that the *majority* of peak traffic is from two *known* sources - Netflix and Youtube. Very well known sources. We *do* know the bitrates that Netflix uses, and we know the bitrates that Youtube uses. We even know that both are buffered significantly by the client, so jitter does not matter for these flows. We know they are pre-precorded, not live, so a delay of even 1000ms or more doesn't matter. We know that alotting more bandwidth than they use would be wasteful and alotting less will make the customer's experience worse. Pretending we don't know this stuff just makes everything worse all around.
On the other hand, you mentioned a flow from "MyFunnyHomeVideos.com". We DON'T know anything about the stream we see as packets sourced from 24.76.120.56. We don't even know if that video stream is a live teleconference, so 1000ms delay would be really bad. For Netflix, we *know* huge delay is fine. For random video stream, the samw delay could be really bad. Treating them the same makes for unhappy customers.
> I'm not aware of any Net Neutrality law that prevents path selection by ISP.
I'm pretty sure that most people who say "all video packets must be treated the same", they would *not* be happy with selecting the "best" link for Netflix and the "worse" link for a no-name video stream from a random source. Maybe they need to say what they mean, but that's difficult because any of the three links is the "best", depending on what you measure.
You say "this is about...". We all know what it's *about*, writing the actual text is very difficult if it's going to a) achieve the objectives and b) not prohibit smart network management.
In theory, you can have the general concept of network neutrality, and also have QoS. Heck, in theory you can have network neutrality and still have a quality *network*, but writing a net neutrality LAW that doesn't seriously damage efforts to provide quality service is very, very difficult. Carrier network is just complicated. For more information with an example or two see:
Let's talk about what the majority of bandwidth *is* for a residential ISP. Netflix. Not "streaming video", Netflix (and Youtube is huge too). We know the source of traffic, and we know which mix of latency, jitter, packet loss, and bandwidth will provide a clean Netflix stream for our customers. We know exactly which bitrate each flow needs, hell we even know how much the CLIENT is buffering, which tells us how much jitter and delay is acceptable, and when a packet is late enough that it should be dropped. So Netflix (and youtube) is *most* of our traffic, and we know a *lot* about it's requirements. Any competent carrier network engineer quickly learns how to set up your queuing strategies, shapers, policing, and routing to optimize customer experience for *most* of your evening traffic, the Netflix traffic, and which parameters work best for Youtube traffic. We don't know the bitrate needed for some other random streaming video and we damn sure don't know how much the client is caching for random streaming video, or if it's caching at all.
You said you want to ban " prioritization". Let's spend 60 seconds to get a clue about that. I've got three connections in rural Arizona, microwave, copper, and satellite. The microwave connection has the most *bandwidth*, it can send the most packets per second. It also drops the most packets - data sent over that link may or may not arrive. The copper is reliable, and packets get there soon, but it has the lowest bandwidth- it can't carry very much data. The satellite connection can carry more data, but each packet takes a long time in transit - it has to go to space and back. So I've got different options, different treatment for different flows. Btw I also have two ways of getting packets to Arizona in the first place - a direct fiber connection with relatively low bandwidth, and a higher bandwidth connection that takes a 500 mile detour through Los Angeles. Again two differentb routes, neither *better* than the other, but with *different* characteristics. Obviously I'm going to send different flows over different links. Which of those links is "preferential"? We want to toss me in jail or fine me if I send the Netflix flow over the "better" link, so tell me which one is "better" so I can avoid using it.
As I said, any competent carrier engineer in the field knows which link will provide the best experience for Youtube, and which will provide the best experience for Netflix. You're asking them to become incompetent. Nay, you want to pass a law forcing them to become incompetent.
> we all know what net neutrality is about: banning the application of arbitrary and unnecessary prioritization of the ISPs own video, VOIP and similar services relative to competing services.
Oh I know what you want. Actually better than you do - you don't know when you want low jitter and when you want low loss. I said I agree with that general concept as a goal. The thing is, while you can set up several different shaping, routing, and policing algorithms in a Cisco router, none of those algorithms is called "you know what I mean, be fair".
I've had to study over 5,000 pages to learn how to choose and configure algorithms for choosing routes, traffic shaping, traffic policing, etc just on Cisco equipment alone. I say "over 5,000", it could easily be 10,000 pages. You seem to be under the impression that none of that science exists, that proper configuration of a carrier network can be defined in a sentence or two.
I'll tell you what, why don't you go spend 15 minutes learning about what some of the most important queuing algorithms are. While you're reading about them, think about how they might apply to a) Netflix, a huge number of large packets with well-known characteristics, b) Youtube, with different characteristics, and c) unknown video, with unknown characteristics. Then we can come back and talk about your series of tubes. See you then.
That sounds good at first, for a second or two, any is a reasonable *general concept*, a one-sentence summary of a 500 page policy.
Let's look at "differential treatment". I've got three connections in rural Arizona, microwave, copper, and satellite. The microwave connection has the most *bandwidth*, it can send the most packets per second. It also drops the most packets - data sent over that link may or may not arrive. The copper is reliable, and packets get there soon, but it has the lowest bandwidth- it can't carry very much data. The satellite connection can carry more data, but each packet takes a long time in transit - it has to go to space and back. So I've got different options, different treatment for different flows. None are "preferential", none *better* than the other, all *different*. Btw I also have two ways of getting packets to Arizona in the first place - a direct fiber connection with relatively low bandwidth, and a higher bandwidth connection that takes a 500 mile detour through Los Angeles. Again two differentb routes, neither *better* than the other, but with *different* characteristics.
Over half of the traffic on a residential ISP comes from one source - Netflix. If you've worked a few days as a carrier network engineer, you know the flow characteristics that Netflix needs - the right balance of bandwidth, delay, jitter, and packet loss that provides your customers a good experienceb with Netflix. You quickly learn how to set up your queuing strategies, shapers, policing, and routing to optimize customer experience for *most* of your evening traffic, the Netflix traffic. Some would pass a law requiring us to pretend to be stupid and ignore all that we know about providing good service, making it illegal to set the shapers, routes, and queues properly knowing what we know about Netflix traffic. That's just really ignorant. The result would be that your experience for *all* flows would get worse, if it's illegal for us to use our knowledge of the requirements of the major services you want to use (Netflix and Youtube mostly, those two are 75% of residential traffic.)
> something that obviously has helped small companies and the internet grow all these years
Huh? We haven't HAD net neutrality regulations "all these years". The FCC rule on network neutrality was issued in mid 2015 and the first enforcement letters sent in the last couple of months. If you think what we've been doing "all of these years" has helped the internet and small companies grow, that's an argument AGAINST Wheeler's new net neutrality regulations.
The argument FOR network neutrality is that ISPs might in the future stop continually improving service and switch to a model that would be bad. That's a legitimate concern, and the intention behind the network neutrality rules was good.
HOWEVER, modern carrier networks are exceedingly complex, and getting more complex all the time. "A packet is a packet is a packet" is a recipe to create horrible service for everyone. Modern are WAY more intelligent than that, and need to be if youb want usable voip that doesn't sound like satellite news coverage, with 1000ms of delay after each thing you say. Laws enforcing network neutrality, if they were written to avoid a lot of unpleasant, unintended consequences, would need to be perhaps 500 pages long. That's *my* issue with Wheeler's regulations - I like general concept, but it was horribly oversimplified, dumbed down to the point of being stupid. A draft rule (not the final rule) would have outlawed blocking spam - you have to treat every source equally means you can't discriminate against spammers. The final rule was *slightly* more nuanced than that, but not by much.
My own opinion is that we should have very specific rules, tailored to objectional behaviors that ISPs are actually doing or about to do, rather than a huge overbroad rule based on a nebulous fear of what some ISP *might* someday do. The overbroad, dumbed down rule criminalizes intelligent network management, in the name of trying to prohibit something that nobody is doing anyway. As an example, one sender, a major mailing list, sends emails to 35,000 of your customers. Then another sender, Bob, sends an email to *one* of your customers, an email from one person to another. It'll take your mail server an hour to churn through the 35,001 emails and deliver them all. Should Bob's person-to-person email sit in the queue for an hour while you first process the 35,000 copies of the "Deal of the Week" email? Intelligent management of your service says that you deprioritize the bulk sender. Is that allowd by Wheeler's rule? Maybe, maybe it'll get you in legal trouble. (That may depend on if the bulk sender is the DNC or not.)
It gets complicated when you get into the technical details of actually implementing it without making service worse for everyone. For that reason, I think we're better off narrowly targeting specific actual problems, rather than Wheeler's shotgun approach.
Distributing child porn, when done by the FBI, may be illegal. I don't feel like reading the statute right now, many laws have exceptions for law enforcement in the course of their duties.
That, however, has nothing whatsoever to do with entrapment. Entrapment is when a person with no intention of committing any crime is induced to do so by the police.
If a person decides of their own free will to go to a child porn site and start downloading videos called "12 year old fucked.wmv" there is no entrapment. They've already decided to download and view that. Whether or not the police track the IP or anything else can't make it entrapment.
What *would* be entrapment would be if an undercover cop pretending to be their friend said to a person: "You know a lot about computers and security and all that, right? You have that Thor thing or whatever? I want to download some stuff without being tracked. I'll give you $50 if you download '12 year old fucked.wmv' for me and put it on a USB drive."
I manage a vulnerable assessment system. We have hundreds of thousands of distinct vulnerabilities in our database, which we look for on the hundreds of thousands of devices we scan every week. I've been working full time in network security for 18 years. Based on the data I have, yes Cisco is *more secure* than most. Especially if the administrator pays attention to security - Cisco provides many, many ways to make your network more secure.
>> Not really, they have bugs too
Anything that has code has bugs. Even most things that DON'T have code have bugs - the average home has more than 100 different kinds of bugs living in it.
It seems perhaps you have some kind of hard-on for criticizing Cisco. That's cool. If you care at all about intellectual honesty, you can point out that Cisco tends to be quite expensive. You could point out that they don't have perfect security. They do definitely do well above average, however, in my experience testing the security of corporate networks.
> Let this be a lesson to companies and non profits a like, its really better to stay out of politics which are beyond your area of direct interest.
In my experience, this is true. I damaged my business by talking about politics on message boards where my customers gather.
On the other hand, I'm a member of a non-profit which has as one of their core principles that they stay out of politics and advance no particular opinion on controversial issues. The organization focuses on their purpose, not getting distracted by the controversy of the month. This has served them well over almost a hundred years.
At least when Congress makes laws, they are doing their job, under their Constitutional authority. Most of the hundreds of thousands of pages of law in the Code of Federal Regulations are unconstitutional - elected Congress reps, who are accountable to the voters entry few years, are supposed to be making law, not unelected, unaccountable bureaucrats.
We have ten times as much regulation today as we did fifty years ago or so. *Some* regulations need to be written by the executive agencies, but many we could well do without and many more are, put simply, laws - and therefore should be passed by Congress after appropriate debate and amendment.
That matches my experience. My company offers an offsite, bootable backup solution so if anything bad happens to your server, you just boot the appropriate clone in our cloud and you're back in business. A LOT of our customers get our service when they find out the hard way why *proper* offsite backups are important. Many weren't too concerned about backup and business continuity until something bad happened to them.
AFTER they have a major loss they get serious about making sure it won't happen again.
The old law is that they must be paid at least 100% of the average. So the company (Infosys, let's name names) plays games to reduce the "average" by 20%, so they can actually pay 20% less than the *true* average.
Now we switch to "whoever pays the highest multiple of the average". As before, Infosys plays games and deflates the average by 20%. My company wants to hire the Linux firewall maintainer, Arturo Borrero Gonzalez. Gonzalez isn't some random programmer, he's the world's foremost expert on Linux firewalls, so my company offers him five times the average programmer salary. What's Infosys going to do?
> So why was WJC not impeached and removed from office for this? Instead we had a trial over presidential blowjobs.
As you may know, two presidents have been impeached - Andrew Johnson and Bill Clinton. The impeachment charges were perjury and obstruction of justice. The Clintons are of course pretty good at obstructing justice, they aren't stupid criminals - Hillary has virtually bragged about that. It's tough to prove the underlying crimes beyond a reasonable doubt, but the obstruction is apparent.
Bill Clinton pled guilty to perjury in state court and was fined $90,000 along with losing his law license for five years.
The masses enjoyed talking about blowjobs, some talked about Hillary's role smearing Paula Jones and others that Bill harassed/had afairs with, but the typical man on the street doesn't sit around talking about obstruction of justice (or emoluments). The trial was about the perjury and obstruction of justice, though, not blowjobs.
An impeachment on violations of the emoluments charge would have been more difficult because of the relative lack of clear law on the matter, and the Clinton's skill at skirting the edges of what legislation does exist. The perjury and obstruction charges were much more clear - again, Bill pled guilty to perjury. He wouldn't have pled guilty to violating the emoluments clause, he would have had much more wiggle room.
I do similar, using LVM and mdadm. I've found it works well. Reliability is much increased by a) automatic monitoring of SMART data which warns me of impending failures via email and b) weekly scrubbing, checking that all blocks are consistent.
> As long as I replace the drives when it fails
The above monitoring and scrubbing lets me replace drives shortly BEFORE they fail, and mostly ensures that the remaining disks don't have hidden errors. A rebuild is intensive, so it can certainly cause a "working" disk to fail at the worst possible time, if you're not verifying the health of those "working" disks weekly.
Google's cloud has some of the same security services that are popular on Amazon, and others can be deployed whenever more customers want them.
As an example, several security services that you can buy through Amazon are actually provided by a security company called Alert Logic. Alert Logic announced support for Google Cloud in 2014:
You can bet that as Google Cloud manages to get more market share, companies like Alert Logic are ready to deploy more services to Google Cloud. They have stuff in the pipeline, awaiting completion whenever their customers ask for Google support.
> how does Google buying Windows licenses pose a threat to Microsoft?
Windows licenses in bulk are what, about $10, once? Versus $2,400 / year for cloud hosting a medium sized instance.
> Wouldn't it be a bigger threat/opportunity to assist companies move away from the Microsoft environment?
It's much easier to sell Google database services, machine learning, DNS, and all the other non-Microsoft Google stuff to companies hosted in Google's cloud. To help companies transition away from Microsoft, Google first needs to allow them to bring their environment to Google. Once it's on Google's cloud platform, it makes more sense for the company's *next* database server to be one of the many databases offered by Google rather than defaulting to MS SQL.
> his company and he's neither sold it nor divested any of the foreign income businesses?
Three of our first four presidents (George Washington, Thomas Jefferson, and James Madison) also had businesses which had foreign customers. None of the founding fathers had a problem with that. If someone wants to buy tobacco from Washington's field (or a hotel room from Trump's hotel) that's not a problem, the founders said.
What they DID have a problem with, and prohibited in the Constitution, was accepting payments not for a product or service, but as a result of the President's position. For example, if a President or Secretary of State received a "gift" of $100,000 from a foreign government, rather than selling a product worth $100,000 THAT'S an unconstitutional emolument. The Supreme Court has confirmed that.
Guess which President and Secretary of State got big "gifts" (emoluments) from from other countries? Hint - they are married, one was president, one was secretary of state. Both took foreign payments just for being in high office, not as part of a routine business transaction selling something.
Slashdot Mirror
> Drop them at random and let the end user see the stupid design decision made by their company of choice.
You could do that, but Netflix or Youtube doesn't work well on your ISP, the customers don't yell at Youtube. And when the server retransmits the packets you dropped at random, it makes the network more congested for *everyone*.
"Just make it worse for everyone" doesn't sound like the best idea to me.
Zero rating does bring up a couple of important issues, agreed.
I think there has been confusion all around. A few years ago, the big controversy around "network neutrality" was basically that Netflx wanted to be the only web site in the world who didn't have to pay their hosting bill. They intentionally confused that with network neutrality and many people on Slashdot, perhaps most, fell for it. That's a *completely* separate issue from anything related to zero rating.
> your stance that carriers ought to be able to do anything they want with the traffic
I've said the exact opposite several times in this thread. I've said I think we need specific rules directed at specific issues.
> why don't you volunteer to write the legal text that would enable the objectives of net neutrality
(and everyone here, including you, knows damn well what those objectives are)
I note you used the plural objectiveS. That's insightful given that other people posting in this thread have referred to the OBJECTIVE, singular, but three different people posting have said three different objectives. If you want to address zero rating, make a rule about zero rating. If peering agreements are a problem, make a rule about that. If providing different services to paid hosting customers than you do for unpaid, unrelated companies is a problem, that's a separate rule.
The issue that's probably most clear to me is if Comcast was artificially slowing Netflix traffic at one point in time. Maybe probably they were - only certain locations experienced the issue, suggesting that the cause *may* have been related to peering, a legitimate business negotiation. It's a lot simpler to address *that* issue, if they were artificially throttling it within Comcast's network, without conflating it with zero rating and other issues.
The discussion of NN heated up, and started appearing on the Slashdot regularly, a few years ago when Netflix, which was smaller at the time, was basically demanding transit service for free. They thought they should be the only web site in the world who doesn't have to pay their web hosting bill. "Everybody here knows", but some here were suckered by Netflix into confusing that peering negotiation with network neutrality. On *that*, Netflix's wish for free peering, I don't think we need any regulation - they can negotiate those arrangements just like every other business in the world does.
A trickier issue is this. Suppose I have a server with some videos on it and I want to stream those to the internet. Obviously I need to connect that server to the internet, through some kind of ISP. So maybe I go to Comcast Business and buy a 100Mbps or gigabit connection for my server(s). That's standard procedure for how you set up a web site. I pay my monthly bill, they sell me a gigabit of connectivity to their network. In real life, I currently have servers in three locations, connected to three ISPs (the B2B side of the ISPs, selling dedicated bandwidth to their network and transit off of their network). Like anyone else hosting a web site of any significant size, there are two numbers that describe how much bandwidth I get, basically the minimum guaranteed and the maximum possible (CIR and signaling rate). I decide how much guaranteed bandwidth I want to buy and what signaling rate I want. The guaranteed bandwidth (CIR) can be guaranteed only because it's prioritized over any non-guaranteed bandwidth. Does that make sense so far? Thr point is that to host servers, you pay for two bandwidth rates - minimum guaranteed and maximum possible. Everybody's minimum bandwidth is prioritized over everybody's non-guaranteed portion. So ...
Suppose Hulu does the exact same thing that everyone with a large web site does and buys an internet connection for one of their servers. They happen to buy this connection from Time Warner. Picking from the price list, they choose 1Gbps CIR (guaranteed) on a 10Gbps connection. That's a normal web hosting contract. But ... fuck! It's also "paid prioritization"! Many people think "paid prioritization" should be illegal, but there's no technical difference between "paid prioritization" and "web hosting". I'm not sure *how* you'd write that law. Do you say that any given web site is only allowed to have servers at one location, so they aren't buying hosting (with traffic guarantees) from several different ISPs? That seems like a bad solution. There's literally no difference between me paying Time Warner to host my one server for raysvideos.com and Netflix paying them to host a cache server. Do *you* have any ideas for an effective rule on that topic?
Someone who has truly mastered their craft may perfection 99% of the time. Or not - Tom Brady completes 64% of his passes.
Suppose the Firefox programmers were the most competent human beings to ever walk the earth, and got it right 99.99% of the time. With 14 million lines of code, they would have 14,000 flaws.
On the other hand, if the Rust string handling functions don't permit buffer overflows, they don't permit buffer overflows - ever. You can't write a buffer overflow in a language that doesn't use buffers. Not only will there not *be* such errors, but you can *prove* there are no such errors, you can trust it.
I don't have any opinion on Rust specifically, good or bad. I'm sure it has tradeoffs. The idea that you shouldn't use reliable tools because humans should just be perfect os silly.
> In practice you can too. Net neutrality is about the source of the data. QoS is about the content. They are very easily distinguished
Okay so I've got some packets from 45.83.129.42. I can tell it's some kind of video. Maybe it's a live teleconference, meaning delay would be really bad and any late packets need to be dropped - they won't be used anyway. Or maybe it's a pre-recorded video and the client is caching 30 seconds, so delay doesn't matter and late packets should be delivered and even retried several times. *We don't know.* How should I treat those packets? When a packet is late should I drop it, forward it, retry it?
I also have some packets from Netflix (actually most of my packets are from Netflix. I *do* know it's not a live teleconference, and latency doesn't matter. I even know how the client-side caching works. How should I treat this flow? Should I pretend I'm stupid and drop late packets, causing the sender to retry them and slow down the network for everyone?
When you make shaping and routing decisions, you can trade bandwidth for packet loss and latency for jitter, on a flow-by-flow basis. I don't need to upgrade anything in order to deliver the right mix of jitter, rate, loss, and latency that works well for Netflix streams. The only reason I wouldn't be able to deliver what customers want is that some of you want to make it illegal for me to do so.
I understand your point, I believe I know what you want.
I think what your missing is that the *majority* of peak traffic is from two *known* sources - Netflix and Youtube. Very well known sources. We *do* know the bitrates that Netflix uses, and we know the bitrates that Youtube uses. We even know that both are buffered significantly by the client, so jitter does not matter for these flows. We know they are pre-precorded, not live, so a delay of even 1000ms or more doesn't matter. We know that alotting more bandwidth than they use would be wasteful and alotting less will make the customer's experience worse. Pretending we don't know this stuff just makes everything worse all around.
On the other hand, you mentioned a flow from "MyFunnyHomeVideos.com". We DON'T know anything about the stream we see as packets sourced from 24.76.120.56. We don't even know if that video stream is a live teleconference, so 1000ms delay would be really bad. For Netflix, we *know* huge delay is fine. For random video stream, the samw delay could be really bad. Treating them the same makes for unhappy customers.
> I'm not aware of any Net Neutrality law that prevents path selection by ISP.
I'm pretty sure that most people who say "all video packets must be treated the same", they would *not* be happy with selecting the "best" link for Netflix and the "worse" link for a no-name video stream from a random source. Maybe they need to say what they mean, but that's difficult because any of the three links is the "best", depending on what you measure.
You say "this is about ...". We all know what it's *about*, writing the actual text is very difficult if it's going to a) achieve the objectives and b) not prohibit smart network management.
In theory, you can have the general concept of network neutrality, and also have QoS. Heck, in theory you can have network neutrality and still have a quality *network*, but writing a net neutrality LAW that doesn't seriously damage efforts to provide quality service is very, very difficult. Carrier network is just complicated. For more information with an example or two see:
https://slashdot.org/comments....
> Email takes almost no bandwidth these days.
Let's talk about what the majority of bandwidth *is* for a residential ISP. Netflix. Not "streaming video", Netflix (and Youtube is huge too). We know the source of traffic, and we know which mix of latency, jitter, packet loss, and bandwidth will provide a clean Netflix stream for our customers. We know exactly which bitrate each flow needs, hell we even know how much the CLIENT is buffering, which tells us how much jitter and delay is acceptable, and when a packet is late enough that it should be dropped. So Netflix (and youtube) is *most* of our traffic, and we know a *lot* about it's requirements. Any competent carrier network engineer quickly learns how to set up your queuing strategies, shapers, policing, and routing to optimize customer experience for *most* of your evening traffic, the Netflix traffic, and which parameters work best for Youtube traffic. We don't know the bitrate needed for some other random streaming video and we damn sure don't know how much the client is caching for random streaming video, or if it's caching at all.
You said you want to ban " prioritization". Let's spend 60 seconds to get a clue about that. I've got three connections in rural Arizona, microwave, copper, and satellite. The microwave connection has the most *bandwidth*, it can send the most packets per second. It also drops the most packets - data sent over that link may or may not arrive. The copper is reliable, and packets get there soon, but it has the lowest bandwidth- it can't carry very much data. The satellite connection can carry more data, but each packet takes a long time in transit - it has to go to space and back. So I've got different options, different treatment for different flows. Btw I also have two ways of getting packets to Arizona in the first place - a direct fiber connection with relatively low bandwidth, and a higher bandwidth connection that takes a 500 mile detour through Los Angeles. Again two differentb routes, neither *better* than the other, but with *different* characteristics. Obviously I'm going to send different flows over different links. Which of those links is "preferential"? We want to toss me in jail or fine me if I send the Netflix flow over the "better" link, so tell me which one is "better" so I can avoid using it.
As I said, any competent carrier engineer in the field knows which link will provide the best experience for Youtube, and which will provide the best experience for Netflix. You're asking them to become incompetent. Nay, you want to pass a law forcing them to become incompetent.
> we all know what net neutrality is about: banning the application of arbitrary and unnecessary prioritization of the ISPs own video, VOIP and similar services relative to competing services.
Oh I know what you want. Actually better than you do - you don't know when you want low jitter and when you want low loss. I said I agree with that general concept as a goal. The thing is, while you can set up several different shaping, routing, and policing algorithms in a Cisco router, none of those algorithms is called "you know what I mean, be fair".
I've had to study over 5,000 pages to learn how to choose and configure algorithms for choosing routes, traffic shaping, traffic policing, etc just on Cisco equipment alone. I say "over 5,000", it could easily be 10,000 pages. You seem to be under the impression that none of that science exists, that proper configuration of a carrier network can be defined in a sentence or two.
I'll tell you what, why don't you go spend 15 minutes learning about what some of the most important queuing algorithms are. While you're reading about them, think about how they might apply to a) Netflix, a huge number of large packets with well-known characteristics, b) Youtube, with different characteristics, and c) unknown video, with unknown characteristics. Then we can come back and talk about your series of tubes. See you then.
That sounds good at first, for a second or two, any is a reasonable *general concept*, a one-sentence summary of a 500 page policy.
Let's look at "differential treatment". I've got three connections in rural Arizona, microwave, copper, and satellite. The microwave connection has the most *bandwidth*, it can send the most packets per second. It also drops the most packets - data sent over that link may or may not arrive. The copper is reliable, and packets get there soon, but it has the lowest bandwidth- it can't carry very much data. The satellite connection can carry more data, but each packet takes a long time in transit - it has to go to space and back. So I've got different options, different treatment for different flows. None are "preferential", none *better* than the other, all *different*. Btw I also have two ways of getting packets to Arizona in the first place - a direct fiber connection with relatively low bandwidth, and a higher bandwidth connection that takes a 500 mile detour through Los Angeles. Again two differentb routes, neither *better* than the other, but with *different* characteristics.
Over half of the traffic on a residential ISP comes from one source - Netflix. If you've worked a few days as a carrier network engineer, you know the flow characteristics that Netflix needs - the right balance of bandwidth, delay, jitter, and packet loss that provides your customers a good experienceb with Netflix. You quickly learn how to set up your queuing strategies, shapers, policing, and routing to optimize customer experience for *most* of your evening traffic, the Netflix traffic. Some would pass a law requiring us to pretend to be stupid and ignore all that we know about providing good service, making it illegal to set the shapers, routes, and queues properly knowing what we know about Netflix traffic. That's just really ignorant. The result would be that your experience for *all* flows would get worse, if it's illegal for us to use our knowledge of the requirements of the major services you want to use (Netflix and Youtube mostly, those two are 75% of residential traffic.)
> something that obviously has helped small companies and the internet grow all these years
Huh? We haven't HAD net neutrality regulations "all these years". The FCC rule on network neutrality was issued in mid 2015 and the first enforcement letters sent in the last couple of months. If you think what we've been doing "all of these years" has helped the internet and small companies grow, that's an argument AGAINST Wheeler's new net neutrality regulations.
The argument FOR network neutrality is that ISPs might in the future stop continually improving service and switch to a model that would be bad. That's a legitimate concern, and the intention behind the network neutrality rules was good.
HOWEVER, modern carrier networks are exceedingly complex, and getting more complex all the time. "A packet is a packet is a packet" is a recipe to create horrible service for everyone. Modern are WAY more intelligent than that, and need to be if youb want usable voip that doesn't sound like satellite news coverage, with 1000ms of delay after each thing you say. Laws enforcing network neutrality, if they were written to avoid a lot of unpleasant, unintended consequences, would need to be perhaps 500 pages long. That's *my* issue with Wheeler's regulations - I like general concept, but it was horribly oversimplified, dumbed down to the point of being stupid. A draft rule (not the final rule) would have outlawed blocking spam - you have to treat every source equally means you can't discriminate against spammers. The final rule was *slightly* more nuanced than that, but not by much.
My own opinion is that we should have very specific rules, tailored to objectional behaviors that ISPs are actually doing or about to do, rather than a huge overbroad rule based on a nebulous fear of what some ISP *might* someday do. The overbroad, dumbed down rule criminalizes intelligent network management, in the name of trying to prohibit something that nobody is doing anyway. As an example, one sender, a major mailing list, sends emails to 35,000 of your customers. Then another sender, Bob, sends an email to *one* of your customers, an email from one person to another. It'll take your mail server an hour to churn through the 35,001 emails and deliver them all. Should Bob's person-to-person email sit in the queue for an hour while you first process the 35,000 copies of the "Deal of the Week" email? Intelligent management of your service says that you deprioritize the bulk sender. Is that allowd by Wheeler's rule? Maybe, maybe it'll get you in legal trouble. (That may depend on if the bulk sender is the DNC or not.)
It gets complicated when you get into the technical details of actually implementing it without making service worse for everyone. For that reason, I think we're better off narrowly targeting specific actual problems, rather than Wheeler's shotgun approach.
We don't need no damn progressive apps in Texas!
They keep showing up, saying they know we have jobs for them here, so we set up a safe space for them, called Austin.
> I have 64GB of internal storage (no sd card slot) and have 67% free space currently
In hope you're enjoying your brand new phone. ;)
Distributing child porn, when done by the FBI, may be illegal. I don't feel like reading the statute right now, many laws have exceptions for law enforcement in the course of their duties.
That, however, has nothing whatsoever to do with entrapment. Entrapment is when a person with no intention of committing any crime is induced to do so by the police.
If a person decides of their own free will to go to a child porn site and start downloading videos called "12 year old fucked.wmv" there is no entrapment. They've already decided to download and view that. Whether or not the police track the IP or anything else can't make it entrapment.
What *would* be entrapment would be if an undercover cop pretending to be their friend said to a person:
"You know a lot about computers and security and all that, right? You have that Thor thing or whatever? I want to download some stuff without being tracked. I'll give you $50 if you download '12 year old fucked.wmv' for me and put it on a USB drive."
THAT would be entrapment.
> So are they are more secure than the next guy?
I manage a vulnerable assessment system. We have hundreds of thousands of distinct vulnerabilities in our database, which we look for on the hundreds of thousands of devices we scan every week. I've been working full time in network security for 18 years. Based on the data I have, yes Cisco is *more secure* than most. Especially if the administrator pays attention to security - Cisco provides many, many ways to make your network more secure.
>> Not really, they have bugs too
Anything that has code has bugs. Even most things that DON'T have code have bugs - the average home has more than 100 different kinds of bugs living in it.
It seems perhaps you have some kind of hard-on for criticizing Cisco. That's cool. If you care at all about intellectual honesty, you can point out that Cisco tends to be quite expensive. You could point out that they don't have perfect security. They do definitely do well above average, however, in my experience testing the security of corporate networks.
> Let this be a lesson to companies and non profits a like, its really better to stay out of politics which are beyond your area of direct interest.
In my experience, this is true. I damaged my business by talking about politics on message boards where my customers gather.
On the other hand, I'm a member of a non-profit which has as one of their core principles that they stay out of politics and advance no particular opinion on controversial issues. The organization focuses on their purpose, not getting distracted by the controversy of the month. This has served them well over almost a hundred years.
I can sure understand that sentiment.
At least when Congress makes laws, they are doing their job, under their Constitutional authority. Most of the hundreds of thousands of pages of law in the Code of Federal Regulations are unconstitutional - elected Congress reps, who are accountable to the voters entry few years, are supposed to be making law, not unelected, unaccountable bureaucrats.
We have ten times as much regulation today as we did fifty years ago or so. *Some* regulations need to be written by the executive agencies, but many we could well do without and many more are, put simply, laws - and therefore should be passed by Congress after appropriate debate and amendment.
That matches my experience. My company offers an offsite, bootable backup solution so if anything bad happens to your server, you just boot the appropriate clone in our cloud and you're back in business. A LOT of our customers get our service when they find out the hard way why *proper* offsite backups are important. Many weren't too concerned about backup and business continuity until something bad happened to them.
AFTER they have a major loss they get serious about making sure it won't happen again.
It' not perfect, for sure.
The old law is that they must be paid at least 100% of the average. So the company (Infosys, let's name names) plays games to reduce the "average" by 20%, so they can actually pay 20% less than the *true* average.
Now we switch to "whoever pays the highest multiple of the average". As before, Infosys plays games and deflates the average by 20%. My company wants to hire the Linux firewall maintainer, Arturo Borrero Gonzalez. Gonzalez isn't some random programmer, he's the world's foremost expert on Linux firewalls, so my company offers him five times the average programmer salary. What's Infosys going to do?
In a word, politics.
> So why was WJC not impeached and removed from office for this? Instead we had a trial over presidential blowjobs.
As you may know, two presidents have been impeached - Andrew Johnson and Bill Clinton. The impeachment charges were perjury and obstruction of justice. The Clintons are of course pretty good at obstructing justice, they aren't stupid criminals - Hillary has virtually bragged about that. It's tough to prove the underlying crimes beyond a reasonable doubt, but the obstruction is apparent.
Bill Clinton pled guilty to perjury in state court and was fined $90,000 along with losing his law license for five years.
The masses enjoyed talking about blowjobs, some talked about Hillary's role smearing Paula Jones and others that Bill harassed/had afairs with, but the typical man on the street doesn't sit around talking about obstruction of justice (or emoluments). The trial was about the perjury and obstruction of justice, though, not blowjobs.
An impeachment on violations of the emoluments charge would have been more difficult because of the relative lack of clear law on the matter, and the Clinton's skill at skirting the edges of what legislation does exist. The perjury and obstruction charges were much more clear - again, Bill pled guilty to perjury. He wouldn't have pled guilty to violating the emoluments clause, he would have had much more wiggle room.
I do similar, using LVM and mdadm. I've found it works well. Reliability is much increased by a) automatic monitoring of SMART data which warns me of impending failures via email and b) weekly scrubbing, checking that all blocks are consistent.
> As long as I replace the drives when it fails
The above monitoring and scrubbing lets me replace drives shortly BEFORE they fail, and mostly ensures that the remaining disks don't have hidden errors. A rebuild is intensive, so it can certainly cause a "working" disk to fail at the worst possible time, if you're not verifying the health of those "working" disks weekly.
Google's cloud has some of the same security services that are popular on Amazon, and others can be deployed whenever more customers want them.
As an example, several security services that you can buy through Amazon are actually provided by a security company called Alert Logic. Alert Logic announced support for Google Cloud in 2014:
https://www.alertlogic.com/pre...
You can bet that as Google Cloud manages to get more market share, companies like Alert Logic are ready to deploy more services to Google Cloud. They have stuff in the pipeline, awaiting completion whenever their customers ask for Google support.
> how does Google buying Windows licenses pose a threat to Microsoft?
Windows licenses in bulk are what, about $10, once? Versus $2,400 / year for cloud hosting a medium sized instance.
> Wouldn't it be a bigger threat/opportunity to assist companies move away from the Microsoft environment?
It's much easier to sell Google database services, machine learning, DNS, and all the other non-Microsoft Google stuff to companies hosted in Google's cloud. To help companies transition away from Microsoft, Google first needs to allow them to bring their environment to Google. Once it's on Google's cloud platform, it makes more sense for the company's *next* database server to be one of the many databases offered by Google rather than defaulting to MS SQL.
> his company and he's neither sold it nor divested any of the foreign income businesses?
Three of our first four presidents (George Washington, Thomas Jefferson, and James Madison) also had businesses which had foreign customers. None of the founding fathers had a problem with that. If someone wants to buy tobacco from Washington's field (or a hotel room from Trump's hotel) that's not a problem, the founders said.
What they DID have a problem with, and prohibited in the Constitution, was accepting payments not for a product or service, but as a result of the President's position. For example, if a President or Secretary of State received a "gift" of $100,000 from a foreign government, rather than selling a product worth $100,000 THAT'S an unconstitutional emolument. The Supreme Court has confirmed that.
Guess which President and Secretary of State got big "gifts" (emoluments) from from other countries? Hint - they are married, one was president, one was secretary of state. Both took foreign payments just for being in high office, not as part of a routine business transaction selling something.