Alex Balashov | 2 Dec 02:20

Re: OT: What do you guys think of this?

RE Kushner List Account wrote:

> The question is, what are you actually paying for as a customer?  To 
> discriminate against bits just because they actually use what they are 
> paying for is beyond me.
> 
> At least a bandwidth cap is easier to understand. You get what you pay for.

Speaking as a former sysadmin of an ISP, I would say that the issue is 
the following:

1) There is a high correlation of network-disrupting levels of traffic 
and BitTorrent;

2) Unlike some "bursty" downloads (like your CentOS ISO from an FTP 
server), BitTorrent traffic has the tendency to be sustained at higher 
levels for longer periods since the architecture presumes that 
everyone's a client and everyone's a server and fragments are always 
moving around.  This is what tends to upset oversubscription assumptions 
that are otherwise functional, and are the only way that the ISP can 
possibly afford to give you the bandwidth for the price of 
consumer-grade broadband.

I would tend to agree with you that discriminating against types of 
services and/or traffic through rate-limiting buckets and deep packet 
inspection is worse than a blanket bandwidth cap.   However, you need to 
keep in mind the other side of the coin;  were it not for Torrent, there 
would not be a need for traffic policing (in the overwhelming 
preponderance of cases) either way, so it's considered unfair to punish 
everyone with a bandwidth cap on everything when in reality, it's not a 
(Continue reading)

BJ Weschke | 2 Dec 02:28

Re: OT: What do you guys think of this?

Alex Balashov wrote:
> RE Kushner List Account wrote:
>
>   
>> The question is, what are you actually paying for as a customer?  To 
>> discriminate against bits just because they actually use what they are 
>> paying for is beyond me.
>>
>> At least a bandwidth cap is easier to understand. You get what you pay for.
>>     
>
> Speaking as a former sysadmin of an ISP, I would say that the issue is 
> the following:
>
> 1) There is a high correlation of network-disrupting levels of traffic 
> and BitTorrent;
>
> 2) Unlike some "bursty" downloads (like your CentOS ISO from an FTP 
> server), BitTorrent traffic has the tendency to be sustained at higher 
> levels for longer periods since the architecture presumes that 
> everyone's a client and everyone's a server and fragments are always 
> moving around.  This is what tends to upset oversubscription assumptions 
> that are otherwise functional, and are the only way that the ISP can 
> possibly afford to give you the bandwidth for the price of 
> consumer-grade broadband.
>
>
> I would tend to agree with you that discriminating against types of 
> services and/or traffic through rate-limiting buckets and deep packet 
> inspection is worse than a blanket bandwidth cap.   However, you need to 
(Continue reading)

Alex Balashov | 2 Dec 02:32

Re: OT: What do you guys think of this?

BJ Weschke wrote:
> Alex Balashov wrote:
>> RE Kushner List Account wrote:
>>
>>   
>>> The question is, what are you actually paying for as a customer?  To 
>>> discriminate against bits just because they actually use what they are 
>>> paying for is beyond me.
>>>
>>> At least a bandwidth cap is easier to understand. You get what you pay for.
>>>     
>> Speaking as a former sysadmin of an ISP, I would say that the issue is 
>> the following:
>>
>> 1) There is a high correlation of network-disrupting levels of traffic 
>> and BitTorrent;
>>
>> 2) Unlike some "bursty" downloads (like your CentOS ISO from an FTP 
>> server), BitTorrent traffic has the tendency to be sustained at higher 
>> levels for longer periods since the architecture presumes that 
>> everyone's a client and everyone's a server and fragments are always 
>> moving around.  This is what tends to upset oversubscription assumptions 
>> that are otherwise functional, and are the only way that the ISP can 
>> possibly afford to give you the bandwidth for the price of 
>> consumer-grade broadband.
>>
>>
>> I would tend to agree with you that discriminating against types of 
>> services and/or traffic through rate-limiting buckets and deep packet 
>> inspection is worse than a blanket bandwidth cap.   However, you need to 
(Continue reading)


Gmane