What really irks me about asking Slashdot about the morality of shaping is that you already know the types of responses you will receive.
The above response was modded to 5:Insightful, which really shows you what agenda Slashdotters have on network access:
Application of QoS/Shaping is not EVIL. Failing to give lower QoS (including shaping) of the 5 percent of users who are damaging the reliability of a service for 95% is ludicrous.
The costs involved in these file sharers is significantly greater than the mean, and it is quite likely that no significant Service Level Agreement is in place for home users.
Many of the arguments about legal liability are fallacious and scaremongering.
Any ISP who provided a contract which said or implied that actual download speeds were equivalent to access speeds would be foolish to say the least. I have never seen such a thing.
In the case that 5% of people are upset with throttling, it's financially in the company's interest to let them leave.
Given that at least half of peer-to-peer traffic
is other peoples' copyrighted material, you have no obligation to assist people to break the law.
Threats of legal action could readily be countered by profiling/capturing the traffic sent through the ISP for legal discovery purposes.
I understand that there are many open source users who would readily make use of peer-to-peer systems. I would suggest that you provide a local mirror for open-source projects, which would limit your upstream costs, and remove the need to use peer-to-peer systems to retrieve open-source updates.
Apart from the issues arising from the prior company's intellectual property (source code, architecture, bugs), rewriting an application
doesn't give you any intellectual property
protection if someone in India or China decides
to do the same thing to you.
So, it is worth seeing if there's a new way of
tackling the task (a business process, a novel
architecture) that can be patented. This would
give the new company some intellectual property
to build their application on.
Of course, it's not even worth thinking about
the problem until you are completely disengaged
from the prior company. That sort of chicken and
egg problem has its own legal ramifications
( if you think up the inventive step while at
the existing company it may be 'work for hire' ).
My personal preference is to go and invent something else. I'd be bored stupid reimplementing someone else's idea of a product.
Each subnet has a/64 allocation so that hosts can arrive and pick their own address in the network with very small chance of collision, even without a server.
This is described in RFC2462: IPv6 Stateless Address Autoconfiguration.
The systems also test the addresses for uniqueness (so there's no birthday problem either). This means IPv6 hosts will typically just start getting an address immediately they are plugged into a network, and on average you have more chance to be struck by lightning than to have your (well distributed) IPv6 address selection collide with another host: It just works.
It also removes artificial boundaries to the size of subnets. You won't have to change your subnet plan because 20 more computers are installed on the 'HR' network for example.
" As an AC pointed out to another post, even if we had IPv6 universally deployed
today, that would not be permitted. We LIKE that NAT breaks end-to-end
networking when we want it to. We LIKE having 64K machines port forwarded behind
a single IP address providing services, when we want to. That isn't going to
change, whatever networking stack we use. "
You're conflating the issue of firewalls with NATs. NATs necessarily have a mapping function which looks like a firewall, but firewalls can block sessions directly and effectively without changing packet formats, and breaking the end-to-end principle
for packets which get through.
With regard to hiding internal topology (which is apparently the only reason
to have all the packets come from one address), RFC3041 Privacy addresses can be
used to hide the identity of hosts, and tunneling protocols (such as Mobile IPv6)
can be used to present a location different to the host's actual topological location.
So the points you've made are not advantages for NATs.
The situation with NATs is so dire that almost all IPv4 based applications
which are being standardized now are made to work around NATs (except for
vanilla client/server TCP applications).
Well, people do want end-to-end real-time communications, and therefore
IPv6 or complicated NAT traversal techniques are required. The advantage
of IPv6 in this case is that it "Just Works" when there's a feed.
We've got one of the dell laptops mentioned here. Had to get the binary video driver from Nvidia for X to work, and the unit weighs about 3.2Kg.
It runs quite nicely, except that the performance is lower when the mains power is removed (Mobile CPU's lower their Hz to save power). here are the bogostats:
Slashdot Mirror
The above response was modded to 5:Insightful, which really shows you what agenda Slashdotters have on network access:
Application of QoS/Shaping is not EVIL. Failing to give lower QoS (including shaping) of the 5 percent of users who are damaging the reliability of a service for 95% is ludicrous.
The costs involved in these file sharers is significantly greater than the mean, and it is quite likely that no significant Service Level Agreement is in place for home users.
Many of the arguments about legal liability are fallacious and scaremongering.
Any ISP who provided a contract which said or implied that actual download speeds were equivalent to access speeds would be foolish to say the least. I have never seen such a thing.
In the case that 5% of people are upset with throttling, it's financially in the company's interest to let them leave.
Given that at least half of peer-to-peer traffic is other peoples' copyrighted material, you have no obligation to assist people to break the law. Threats of legal action could readily be countered by profiling/capturing the traffic sent through the ISP for legal discovery purposes.
I understand that there are many open source users who would readily make use of peer-to-peer systems. I would suggest that you provide a local mirror for open-source projects, which would limit your upstream costs, and remove the need to use peer-to-peer systems to retrieve open-source updates.
So, it is worth seeing if there's a new way of tackling the task (a business process, a novel architecture) that can be patented. This would give the new company some intellectual property to build their application on.
Of course, it's not even worth thinking about the problem until you are completely disengaged from the prior company. That sort of chicken and egg problem has its own legal ramifications ( if you think up the inventive step while at the existing company it may be 'work for hire' ).
My personal preference is to go and invent something else. I'd be bored stupid reimplementing someone else's idea of a product.
It's nothing to do with DNS.
The DNS names won't change, they will just be mapped to an IPv6 address using an AAAA record.
Each subnet has a /64 allocation so that hosts can arrive and pick their
own address in the network with very small chance of collision, even
without a server.
This is described in RFC2462: IPv6 Stateless Address Autoconfiguration.
The systems also test the addresses for uniqueness (so there's no
birthday problem either). This means IPv6 hosts will typically just
start getting an address immediately they are plugged into a network,
and on average you have more chance to be struck by lightning than
to have your (well distributed) IPv6 address selection collide
with another host: It just works.
It also removes artificial boundaries to the size of subnets.
You won't have to change your subnet plan because 20 more
computers are installed on the 'HR' network for example.
You don't have to, even if there's no routing or DNS.
n gwg-icmp-name-lookups-15.txt
http://www.ietf.org/internet-drafts/draft-ietf-ip
" As an AC pointed out to another post, even if we had IPv6 universally deployed today, that would not be permitted. We LIKE that NAT breaks end-to-end networking when we want it to. We LIKE having 64K machines port forwarded behind a single IP address providing services, when we want to. That isn't going to change, whatever networking stack we use. " You're conflating the issue of firewalls with NATs. NATs necessarily have a mapping function which looks like a firewall, but firewalls can block sessions directly and effectively without changing packet formats, and breaking the end-to-end principle for packets which get through. With regard to hiding internal topology (which is apparently the only reason to have all the packets come from one address), RFC3041 Privacy addresses can be used to hide the identity of hosts, and tunneling protocols (such as Mobile IPv6) can be used to present a location different to the host's actual topological location. So the points you've made are not advantages for NATs. The situation with NATs is so dire that almost all IPv4 based applications which are being standardized now are made to work around NATs (except for vanilla client/server TCP applications). Well, people do want end-to-end real-time communications, and therefore IPv6 or complicated NAT traversal techniques are required. The advantage of IPv6 in this case is that it "Just Works" when there's a feed.
I think that the issue is not just
the cost of rollout but the cost of not
rolling out.
Sure the 'States can hold out with
IPv4 for the next ten years, but then what?
It will be importing all of the networking
expertise from places like China, Korea and
Japan, who already have experience with
running IPv6 networks.
You'd be kissing the golden age of
technology leadership goodbye.
It may not happen that way, but if we're
talking about business risks, this is one.
Well, Oz can always claim it if it's successful.
remove mains power:
/proc/cpuinfo
/proc/cpuinfo
boot linux
cat
telinit 0
plug in mains power:
boot linux
cat
telinit 0
We've got one of the dell laptops mentioned here.
Had to get the binary video driver from Nvidia for
X to work, and the unit weighs about 3.2Kg.
It runs quite nicely, except that the performance is lower when the mains power is removed (Mobile CPU's lower their Hz to save power).
here are the bogostats:
2385.51 bogomips - 1196.502 MHz (batteries)
3185.04 bogomips - 1595.321 MHz
We're not running quake on it tho'