how-to block ads
|
|
Share Topic
 |
 |
|
|
 GaffEvery Villain Is Lemons
join:1999-09-05
North TX, US | reply to pnh102
Re: Not Real Data ISPs will not know who has the boxes and who doesn't.
--
My PC Gaming Blog
»thegaffadin.blogspot.com | |
| Agreed. I see no reason why an ISP would know who has boxes. In fact, that would be kind of the whole point... | | |
|
 pnh102Reptiles Are Cuddly And PrettyPremium
join:2002-05-02
Mount Airy, MD | reply to Gaff
said by Gaff:ISPs will not know who has the boxes and who doesn't. And again, what is to stop ISPs from simply giving the highest priority to the FCC's testing site, regardless of who has a box or not?
Anyone who has even the most basic understanding of networking knows that it is trivially easy to give the highest traffic priority to a certain target network.
--
"Net Neutrality" zealots - the people you can thank for your capped Internet service. | |
lesopp
join:2001-06-27
Land O Lakes, FL
1 edit | reply to Gaff
The ISP doesn't need to know who has the boxes, but for accuracy the FCC does needs to know the speed test destination. With that in mind, an ISP can get a copy of the application, profile the communication behaviour, then construct and apply priority filters to skew the results in favor of the ISP. | |
 tubbynetreminds me of the danse russePremium,MVM
join:2008-01-16
Chandler, AZ | reply to pnh102
said by pnh102:Anyone who has even the most basic understanding of networking knows that it is trivially easy to give the highest traffic priority to a certain target network. two things.
(a) anyone familiar with networking is also aware that it is more than just your isp determining speeds. if an isp prioritizes by destination network, then that is great - until you leave your isp network to a backbone carrier.
(b) additionally, you're assuming that all gear can perform an ingress policy inspection/marking without having a significant effect on packet throughput or route processor cpu cycles. many carrier class hardware based switches cannot perform packet marking that fast because the architecture was chosen for hardware based input/output, not ability for wirespeed throughput *while* marking. all carrier class gear is *not* created equal.
of course, i assume you think *everyone* should know that 
q.
--
"...if I in my north room dance naked, grotesquely before my mirror waving my shirt round my head and singing softly to myself..." | |
pnh102Reptiles Are Cuddly And PrettyPremium
join:2002-05-02
Mount Airy, MD | Doesn't matter. We still know that ISPs can, and routinely do, give high priority to speed testing sites, so regardless of how they do it, the point still is that it can be done.
--
"Net Neutrality" zealots - the people you can thank for your capped Internet service. | |
tubbynetreminds me of the danse russePremium,MVM
join:2008-01-16
Chandler, AZ | said by pnh102:Doesn't matter. We still know that ISPs can, and routinely do, give high priority to speed testing sites, so regardless of how they do it, the point still is that it can be done. no, it *absolutely* matters -- something you have ignored. what i'm saying is that an isp may not even be able to due to hardware constraints and even if it can be done, if its not within their own network, then the classification and markings aren't even honored. additionally, with the plethora of testing sites available, its just easier for the isp to deliver the speeds promised.
ironic that you're pro business *and* know the ways to screw the consumer.
q.
--
"...if I in my north room dance naked, grotesquely before my mirror waving my shirt round my head and singing softly to myself..." | |
fiberguyMy views are my own.Premium
join:2005-05-20
kudos:3 | reply to pnh102
... sooooooo. great. We'll waste money, in some form or another, to find out what the "average speed in America is" to some one provider or site then eh?
Very nice..
This would assume that all traffic on the internet goes to one place, and it doesn't. This also assumes that everyone is going to take the same path TO that destination over the same backbones as well.
This is just silly.. why, if the GOVERNMENT is so up in arms about this, why don't they simply make up some rules, like they love to do, and invade the ISP AND the homes of customers, including their lives as well, and install their monitoring equipment in THOSE locations.
In the end, the only speed that the ISP can guarantee, or even try to guarantee, is that between the customer and the central office or head end. I don't think there is an argument that the ISPs have enough bandwidth out to the cloud, otherwise their entire city or service area would be consistently slow. Even then, if they find "slow connections" I'd think they'd have to monitor the entire node, in the case of cable, to see if it's an isolated issue or not.. I mean, it just seems like they're trying to find a needle in a haystack becuase there could be any number of reasons why an internet connection slows down - and "averages" don't prove anything either.. like I said, TOO many things that CAN go wrong which include issues INSIDE the home. (Hell, a few weeks ago, I could have SWORE that I was having modem problems - even swapped it out. It worked great for 2 minutes, and then it died again - turned out, one of my boxes had a virus on it and was taking the connection down.. I supposed that was Comcast's fault? .. an average consumer would seem to think so.. been there, seen that, done it.)
But really.. if they want to measure speeds, then they need to treat it like anything else.. go in, be upfront, and monitor the sources directly.. this just all sounds like a 5th grader sneaking around trying to find answers on their own.. its stupid, silly, horribly inaccurate and won't yield anything.
Seriously - they should approach this like weights and measures IF this is so important to them.. | |
fiberguyMy views are my own.Premium
join:2005-05-20
kudos:3 | reply to tubbynet
"ironic that you're pro business *and* know the ways to screw the consumer."
What's wrong? can't handle the truth? What he's telling you is absolutely true and technically capable.. so what's your problem/damage? He's pointing out a FACT to you and you want to ignore it like the three monkeys (see no evil, hear no evil, speak no evil) in the process.
He's telling you that the process and method in which they want to "protect the consumer" is not going to accomplish anything when an ISP can easily manipulate the method.. what don't you get?
.. but since he knows how to "screw the consumer" just simply means on a technical level he understands... do you? I doubt it, or you'd not come back with that reply - at all!
One last thing and this gets old.. just because someone can see a view on another side of things, unlike people like you, doesn't make you "pro business"... it shows a level of intelligence.. it doesn't make you ANTI-CONSUMER or "screw the consumer" as people like you love to say (which is a sign of a narrow minded view) SOMETIMES in order to be pro-consumer, you have to be pro-business.. NOTHING in life comes from nothing.. everything in life comes from something. If you expect a business to BE there to provide a service, then it CAN'T ALWAYS be about "the consumer" as much as the consumer would like it to be.. SOMETIMES the consumer is just a bit off being level in their thinking.. ANY consumer that doesn't understand how a business HAS to run has no place in the discussion.. period. | |
tubbynetreminds me of the danse russePremium,MVM
join:2008-01-16
Chandler, AZ | said by fiberguy:What he's telling you is absolutely true and technically capable.. so what's your problem/damage? He's pointing out a FACT to you and you want to ignore it like the three monkeys (see no evil, hear no evil, speak no evil) in the process. apparently, because i have espoused views that don't line up with yours, my points are automatically invalid.
i have not denied that it is not *technically* possible. from a theoretical perspective, very little is actually *impossible*. however, in practice, theory goes out the window. to fully understand the implications involved, you need to understand the architecture behind each device. there is a very large fundamental difference between hardware and software-based platforms and how they can handle large loads of ingress sorting and policing and tagging with qos/cos values. depending on the interconnection that the cmts/vrad has with the core of the isp network, it may or may not be technically feasible to actually do what pnh102  is suggesting. case in point - the cisco catalyst 6509e with vs-720-3cxl is a hardware based sw/routing platform capable of running cisco monolithic ios. retail, the chassis and the supervisor would run you roughly $35k -- no line cards. that box will happily pass 10gig streams all day long without thinking. easiest way to dos that box (supposing there is no hardware rate limiters, copp, etc) is to run several kpps of icmp traceroute traffic through it. that $35k box just became the loudest paperweight you'll own until you shut off the icmp streams. the reason -- it has to perform software punting from the linecards to the supervisor, a similar process that must be gone through with classification and marking on ingress (if this box is the direct peer with your cmts/vrad unit).
i've also pointed out that *unless* the speedtest is locally hosted (i.e. on your providers network), any classification/tagging/prioritization in queue is lost when you peer to your upstream. so if my isp prioritizes to samknows tests, but i have to go through level(3) to get there, what good is my traffic prioritization within cox?
.. but since he knows how to "screw the consumer" just simply means on a technical level he understands... do you? I doubt it, or you'd not come back with that reply - at all!
really? i'd encourage you to look through the technical aspects of my posting history, then we'll talk.
One last thing and this gets old.. just because someone can see a view on another side of things, unlike people like you, doesn't make you "pro business"... it shows a level of intelligence.. it doesn't make you ANTI-CONSUMER or "screw the consumer" as people like you love to say (which is a sign of a narrow minded view) SOMETIMES in order to be pro-consumer, you have to be pro-business.. NOTHING in life comes from nothing.. everything in life comes from something. If you expect a business to BE there to provide a service, then it CAN'T ALWAYS be about "the consumer" as much as the consumer would like it to be.. SOMETIMES the consumer is just a bit off being level in their thinking.. ANY consumer that doesn't understand how a business HAS to run has no place in the discussion.. period.
again, look through my posting history. as it comes to network congestion avoidance, traffic management, and isp policies, i have sided with the corporation more than the consumer. my viewpoint is this -- government oversight is not there to "create a nanny state". consumers should be given the most complete picture of their purchasing decisions as possible, and government is there to facilitate that. rather than let shady practices and lack of oversight hide what is really bad with company "a" as opposed to "b", i'd rather that the government mandate certain "honesty" in services provided and let the customer make the choice that suits them.
i see no problem with comcasts 250gig cap, cox's protocol specific congestion management system, cell-phone carrier etfs, etc. these are things that have been disclosed in the forefront. if i am forced to sign a contract, i must make sure that i know everything as best as possible before signing something, much like i feel everyone else should do.
now, if you have anything else other than invoking the "american way" and insulting my (assumed) lack of technical *and* business knowledge, i'd love to hear it...
q.
--
"...if I in my north room dance naked, grotesquely before my mirror waving my shirt round my head and singing softly to myself..." | |
fiberguyMy views are my own.Premium
join:2005-05-20
kudos:3 | Naaa.. I think we've insulted each other enough. 
Besides, I guess you and I do agree on many things than not. lol | |
|
