🚀 go-pugleaf

In article <1mqbsi.ap2.17.1@news.alt.net>, Chris Caputo says... 

> Do folks think it would be a useful thing if Altopia provided SSL access 
> at no additional cost?
> 
> Existing customers: Do you want it?
> 
> Non-customers: Would it increase your chances of joining us?
> 
> Thanks for your input.

I can't say that's very high on my priority list, Chris

Moo-oo-oo-oo!!!1!

-- 

"I never thought there were corners in time
 until I was told to go stand in one"
....Grace Slick

Chris Caputo wrote:
> Do folks think it would be a useful thing if Altopia provided SSL access 
> at no additional cost?

For what reason? So you can't get captured downloading nasty spam from 
usenet?

:-)

Uni


> 
> Existing customers: Do you want it?
> 
> Non-customers: Would it increase your chances of joining us?
> 
> Thanks for your input.
> 
> Chris Caputo
> President, Altopia Corporation

Uni <no.email@no.email.invalid> writes:
>Chris Caputo wrote:
>> Do folks think it would be a useful thing if Altopia provided SSL access 
>> at no additional cost?
>
>For what reason? So you can't get captured downloading nasty spam from 
>usenet?

Reasons may vary.  Off the top of my head, it prevents login passwords 
from being sent in plain-text, plus it gives dissidents or anyone else 
added protection from nosy ISPs/backbones/governments with sniffing 
software.  As long as people respect the rights of others, what people 
read or post is their own business.

Chris Caputo
President, Altopia Corporation

Gary L. Burnore wrote:
> On Sun, 09 Mar 2008 03:58:52 GMT, Uni <no.email@no.email.invalid>
> wrote:
> 
> 
>>Chris Caputo wrote:
>>
>>>Do folks think it would be a useful thing if Altopia provided SSL access 
>>>at no additional cost?
>>
>>For what reason? So you can't get captured downloading nasty spam from 
>>usenet?
> 
> 
> Poor unitard.  Just because you've never had a thought doesn't mean
> others haven't.

How's those three subscribers like Datbasix? Whoops!
How's that web host plan coming? Whoops!
Seems you best give it some "thought" how you're make a living, rather 
than sit on your behind!

:-)

Uni


> 

On Sat, 08 Mar 2008 09:30:24 +0000, Chris Caputo wrote:

> Do folks think it would be a useful thing if Altopia provided SSL access
> at no additional cost?...

The last time I checked, the NSP's who were offering SSL were charging
extra for it, and quite a bit.  I thought at the time that they had no
real reason for charging extra, but of course I'm not an NSP.

How much extra cost would you incur for offering SSL access?  If little
or none, then why not just offer it by default?  If it's not trivial,
then where do the extra costs come from?

In message <fr15r7$mfm$1@blackhelicopter.databasix.com> Gary L. Burnore
<gburnore@databasix.com> wrote:

>The cost of an SSL Cert has come down considerably.  As with Email
>SSL, The extra costs are in support.  Getting people set up who've
>never used it, explaining why certain apps don't support it properly,
>etc.

The certificate cost is trivial, the cost is normally the hardware
dedicated to encrypting outbound content to maintain speeds at a
semi-reasonable rate.

engineer2 <engineer2@the.lab> writes:
>On Sat, 08 Mar 2008 09:30:24 +0000, Chris Caputo wrote:
>> Do folks think it would be a useful thing if Altopia provided SSL access
>> at no additional cost?...
>
>The last time I checked, the NSP's who were offering SSL were charging 
>extra for it, and quite a bit.  I thought at the time that they had no 
>real reason for charging extra, but of course I'm not an NSP.
>
>How much extra cost would you incur for offering SSL access?  If little 
>or none, then why not just offer it by default?  If it's not trivial, 
>then where do the extra costs come from?

If Altopia offering SSL at no additional cost results in even just a few 
hundred additional customers, that would quickly pay for additional CPU 
horsepower.  Dual-proc quad-core servers are pretty darn cheap these days.

Regarding comments in other messages about the SSL cert, in a beta test we 
are simply using our web SSL cert and it is working fine.

Now that we've got our multi-part retention up to a sane level (8+ days) 
and text above 300 days, I'm finding it hard to not offer SSL at no 
additional cost, as the logical next step.

So how about it, would you sign up with Altopia if we added this perk?  
Our base offering is $6 per month, throttled but with unlimited 
usage/access.

Chris Caputo
President, Altopia Corporation

Gary L. Burnore wrote:
> On Sun, 09 Mar 2008 12:16:44 -0600, DevilsPGD
> <spam_narf_spam@crazyhat.net> wrote:
> 
> 
>>In message <fr15r7$mfm$1@blackhelicopter.databasix.com> Gary L. Burnore
>><gburnore@databasix.com> wrote:
>>
>>
>>>The cost of an SSL Cert has come down considerably.  As with Email
>>>SSL, The extra costs are in support.  Getting people set up who've
>>>never used it, explaining why certain apps don't support it properly,
>>>etc.
>>
>>The certificate cost is trivial, 
> 
> 
> Gee. Thought I just said that.

Try English the next time.

:-)

Uni

> 
> 
>>the cost is normally the hardware dedicated to encrypting outbound content to maintain speeds at a
>>semi-reasonable rate.
> 
> 
> The hardware cost is also trivial.   

Chris Caputo wrote:
 
>Non-customers: Would it increase your chances of joining us?

Never in a million years, chris.
Prices for a subscription are ridiculous and unattached from reality.

Feel free to get acquainted with the plans that other services are
offering and join the rest of us, on planet Earth in the 21st Century.
>
>Thanks for your input.
>
>Chris Caputo
>President, Altopia Corporation

thorn <thorn@usenet.> writes:
>Chris Caputo wrote:
>>Non-customers: Would it increase your chances of joining us?
>
>Never in a million years, chris.
>Prices for a subscription are ridiculous and unattached from reality.

How so?  $6/month is ridiculous?

>Feel free to get acquainted with the plans that other services are
>offering and join the rest of us, on planet Earth in the 21st Century.

Unless you are getting a better deal than what is offered on your 
provider's web site, it looks like our offerings are less expensive.

Chris Caputo
President, Altopia Corporation

In message <1n19p7.9e8.19.1@news.alt.net> Chris Caputo <ccaputo@alt.net>
wrote:

>thorn <thorn@usenet.> writes:
>>Chris Caputo wrote:
>>>Non-customers: Would it increase your chances of joining us?
>>
>>Never in a million years, chris.
>>Prices for a subscription are ridiculous and unattached from reality.
>
>How so?  $6/month is ridiculous?

You have what, 8-9 days of binary retention?  Compare that to the 100+
available on usenetserver, and you'll understand.

If you're only marketing to text users or low volume binary users,
you're competing with Supernews at ~$3.58/month (if prepaid for one
year, with access to over 700 days of retention for this very group) or
APN at $2.95/month (300+ days of text retention)

Maybe not ridiculous, but around double what the competition offers.

Chris Caputo wrote:
> Uni <no.email@no.email.invalid> writes:
> 
>>Chris Caputo wrote:
>>
>>>Do folks think it would be a useful thing if Altopia provided SSL access 
>>>at no additional cost?
>>
>>For what reason? So you can't get captured downloading nasty spam from 
>>usenet?
> 
> 
> Reasons may vary.  Off the top of my head, it prevents login passwords 
> from being sent in plain-text, plus it gives dissidents or anyone else 
> added protection from nosy ISPs/backbones/governments with sniffing 
> software.  As long as people respect the rights of others, what people 
> read or post is their own business.

So, can you provide me of ANY example where this has been a problem? If 
no, why offer it?

Uni


> 
> Chris Caputo
> President, Altopia Corporation

Chris Caputo wrote:
 
>thorn <thorn@usenet.> writes:
>>Chris Caputo wrote:
>>>Non-customers: Would it increase your chances of joining us?
>>
>>Never in a million years, chris.
>>Prices for a subscription are ridiculous and unattached from reality.
>
>How so?  $6/month is ridiculous?

128,000bps/ 8.8 days of retention is a complete joke for six bucks.
Bet the number of subscribers to this plan can be counted on your
twelve toes.
>
>>Feel free to get acquainted with the plans that other services are
>>offering and join the rest of us, on planet Earth in the 21st Century.
>
>Unless you are getting a better deal than what is offered on your 
>provider's web site, it looks like our offerings are less expensive.

Au Contraire, oh peanut counter.

Access is granted to the unwashed masses for a paltry sum of ~13.00 -
20.00 a month from numerous sources. And through the Grace Of God,
manage to deliver a steady newsfeed that saturates a 15meg connection
whenever the user decides to logon after paying such an idiotic fee.

Then there are prices for those who just fell off a banana boat:
    128,000 bits per second max - US$6.00/month
--
    2,048,000 bits per second max - US$96.00/month

Which equates to ~720.00 a month to receive the same level of service
that is delivered for 15.00 from this provider.



	You must have rocks in yer head, sonny boy. 
>
>Chris Caputo
>President, Altopia Corporation

In message <47D5F0AC.1090008@no.email.invalid> Uni
<no.email@no.email.invalid> wrote:

>Chris Caputo wrote:
>> Uni <no.email@no.email.invalid> writes:
>> 
>>>Chris Caputo wrote:
>>>
>>>>Do folks think it would be a useful thing if Altopia provided SSL access 
>>>>at no additional cost?
>>>
>>>For what reason? So you can't get captured downloading nasty spam from 
>>>usenet?
>> 
>> 
>> Reasons may vary.  Off the top of my head, it prevents login passwords 
>> from being sent in plain-text, plus it gives dissidents or anyone else 
>> added protection from nosy ISPs/backbones/governments with sniffing 
>> software.  As long as people respect the rights of others, what people 
>> read or post is their own business.
>
>So, can you provide me of ANY example where this has been a problem? 

Do you own a laptop?  If so, do you use it on networks that you don't
trust?

As a traveler, when visiting hotels, using hotspots, or one of the
worst, tech trade shows, any plain-text auth credentials are very likely
to be sniffed on a regular basis.

NNTP isn't a huge target, but it's not off the radar either.

>If no, why offer it?

Customer demand.

thorn <thorn@usenet.> writes:
>Chris Caputo wrote:
>>thorn <thorn@usenet.> writes:
>>>Chris Caputo wrote:
>>>>Non-customers: Would it increase your chances of joining us?
>>>
>>>Never in a million years, chris.
>>>Prices for a subscription are ridiculous and unattached from reality.
>>
>>How so?  $6/month is ridiculous?
>
>128,000bps/ 8.8 days of retention is a complete joke for six bucks.
>Bet the number of subscribers to this plan can be counted on your
>twelve toes.
>>
>>>Feel free to get acquainted with the plans that other services are
>>>offering and join the rest of us, on planet Earth in the 21st Century.
>>
>>Unless you are getting a better deal than what is offered on your 
>>provider's web site, it looks like our offerings are less expensive.
>
>Au Contraire, oh peanut counter.
>
>Access is granted to the unwashed masses for a paltry sum of ~13.00 -
>20.00 a month from numerous sources. And through the Grace Of God,
>manage to deliver a steady newsfeed that saturates a 15meg connection
>whenever the user decides to logon after paying such an idiotic fee.
>
>Then there are prices for those who just fell off a banana boat:
>    128,000 bits per second max - US$6.00/month
>--
>    2,048,000 bits per second max - US$96.00/month
>
>Which equates to ~720.00 a month to receive the same level of service
>that is delivered for 15.00 from this provider.

I sincerely appreciate the feedback.

Taking into account our lower retention, if we offered $12/month 
unthrottled, with SSL and 10 streams, would that be reasonable?

Thanks,
Chris Caputo
President, Altopia Corporation

Chris Caputo wrote:
 
>thorn <thorn@usenet.> writes:
>>Chris Caputo wrote:
>>>thorn <thorn@usenet.> writes:
>>>>Chris Caputo wrote:
>>>>>Non-customers: Would it increase your chances of joining us?
>>>>
>>>>Never in a million years, chris.
>>>>Prices for a subscription are ridiculous and unattached from reality.
>>>
>>>How so?  $6/month is ridiculous?
>>
>>128,000bps/ 8.8 days of retention is a complete joke for six bucks.
>>Bet the number of subscribers to this plan can be counted on your
>>twelve toes.
>>>
>>>>Feel free to get acquainted with the plans that other services are
>>>>offering and join the rest of us, on planet Earth in the 21st Century.
>>>
>>>Unless you are getting a better deal than what is offered on your 
>>>provider's web site, it looks like our offerings are less expensive.
>>
>>Au Contraire, oh peanut counter.
>>
>>Access is granted to the unwashed masses for a paltry sum of ~13.00 -
>>20.00 a month from numerous sources. And through the Grace Of God,
>>manage to deliver a steady newsfeed that saturates a 15meg connection
>>whenever the user decides to logon after paying such an idiotic fee.
>>
>>Then there are prices for those who just fell off a banana boat:
>>    128,000 bits per second max - US$6.00/month
>>--
>>    2,048,000 bits per second max - US$96.00/month
>>
>>Which equates to ~720.00 a month to receive the same level of service
>>that is delivered for 15.00 from this provider.
>
>I sincerely appreciate the feedback.
>
>Taking into account our lower retention, if we offered $12/month 
>unthrottled, with SSL and 10 streams, would that be reasonable?



			It's an Epiphany! 


From: ccaputo@alt.net (Chris Caputo)
Subject: Re: Altopia and this group (was: Easynews will filter
misplaced binaries!)
Date: 8 Aug 2002 22:32:37 GMT
Message-ID: <aiuri5$kpc$0@207.14.113.2>

thorn <thorn@use.net> writes:
>Chris Caputo wrote:
>>So all you ex-Altopia customers out there, please come home!!!  Things are
>>better and getting even more so. 
>
>	I am more than willing to "Come Back Home" but your rates are
>ridiculous for subscribers with a cable connection. Do you have any
>plans to make an adjustment of the subscription fees? 

Yes.  In the coming months we will offer unthrottled usage based
accounts.

Chris Caputo
President, Altopia Corporation
>
>Thanks,
>Chris Caputo
>President, Altopia Corporation

thorn <thorn@usenet.> writes:
>Chris Caputo wrote:
>>thorn <thorn@usenet.> writes:
>>>Chris Caputo wrote:
>>>>thorn <thorn@usenet.> writes:
>>>>>Chris Caputo wrote:
>>>>>>Non-customers: Would it increase your chances of joining us?
>>>>>
>>>>>Never in a million years, chris.
>>>>>Prices for a subscription are ridiculous and unattached from reality.
>>>>
>>>>How so?  $6/month is ridiculous?
>>>
>>>128,000bps/ 8.8 days of retention is a complete joke for six bucks.
>>>Bet the number of subscribers to this plan can be counted on your
>>>twelve toes.
>>>>
>>>>>Feel free to get acquainted with the plans that other services are
>>>>>offering and join the rest of us, on planet Earth in the 21st Century.
>>>>
>>>>Unless you are getting a better deal than what is offered on your 
>>>>provider's web site, it looks like our offerings are less expensive.
>>>
>>>Au Contraire, oh peanut counter.
>>>
>>>Access is granted to the unwashed masses for a paltry sum of ~13.00 -
>>>20.00 a month from numerous sources. And through the Grace Of God,
>>>manage to deliver a steady newsfeed that saturates a 15meg connection
>>>whenever the user decides to logon after paying such an idiotic fee.
>>>
>>>Then there are prices for those who just fell off a banana boat:
>>>    128,000 bits per second max - US$6.00/month
>>>--
>>>    2,048,000 bits per second max - US$96.00/month
>>>
>>>Which equates to ~720.00 a month to receive the same level of service
>>>that is delivered for 15.00 from this provider.
>>
>>I sincerely appreciate the feedback.
>>
>>Taking into account our lower retention, if we offered $12/month 
>>unthrottled, with SSL and 10 streams, would that be reasonable?
>
>			It's an Epiphany! 
>
>From: ccaputo@alt.net (Chris Caputo)
>Subject: Re: Altopia and this group (was: Easynews will filter
>misplaced binaries!)
>Date: 8 Aug 2002 22:32:37 GMT
>Message-ID: <aiuri5$kpc$0@207.14.113.2>
>
>thorn <thorn@use.net> writes:
>>Chris Caputo wrote:
>>>So all you ex-Altopia customers out there, please come home!!!  Things are
>>>better and getting even more so. 
>>
>>	I am more than willing to "Come Back Home" but your rates are
>>ridiculous for subscribers with a cable connection. Do you have any
>>plans to make an adjustment of the subscription fees? 
>
>Yes.  In the coming months we will offer unthrottled usage based
>accounts.
>
>Chris Caputo
>President, Altopia Corporation
>>
>>Thanks,
>>Chris Caputo
>>President, Altopia Corporation

My how time flies!  :-)

So, tis good?

Chris Caputo
President, Altopia Corporation

On Wed, 12 Mar 2008 05:25:40 +0000, Chris Caputo wrote:

...

> Taking into account our lower retention, if we offered $12/month
> unthrottled, with SSL and 10 streams, would that be reasonable?

I've noticed that the number of simultaneous connections (streams?)
seems to be important to people, but I've never understood the reason.
Does each one of your simultaneous connections add another increment
of throttled bandwidth?  Other than that, I can't see any advantage.

In another post you mentioned plans for usage-based accounts. By that
do you mean non-expiring blocks?  ATM that's the only kind of account
that I would consider with any NSP.

I've read that even ISP's are considering usage-based accounts (to
penalize you binaries junkies who download 24/7 from unthrottled
NSP's ;-)

engineer2 <engineer2@the.lab> writes:
>On Wed, 12 Mar 2008 05:25:40 +0000, Chris Caputo wrote:
>> Taking into account our lower retention, if we offered $12/month
>> unthrottled, with SSL and 10 streams, would that be reasonable?
>
>I've noticed that the number of simultaneous connections (streams?)
>seems to be important to people, but I've never understood the reason.

I imagine the reasons vary.

With multiple connections you can download binaries in one group while 
browsing text messages in another, for example.

>Does each one of your simultaneous connections add another increment
>of throttled bandwidth?  Other than that, I can't see any advantage.

On our throttled connections, although our terms indicate that bandwidth 
is distributed across parallel connections, I haven't found it necessary 
to actually implement that and enforce it.  Therefore, on our throttled 
offerings, multiple connections will actually increase overall throughput.

In general, with TCP performance testing, I find that a single unthrottled 
stream is best able to utilize available bandwidth.  Others may disagree 
or have more scientific input.

>In another post you mentioned plans for usage-based accounts. By that
>do you mean non-expiring blocks?  ATM that's the only kind of account
>that I would consider with any NSP.

I used to think that usage-based was a good direction but now with the 
growth of flat-rate unlimited accounts, I'm not so sure.  The accounting 
is certainly simpler when flat-rate.

Non-expiring blocks seems unattractive, from the standpoint of a provider 
(this one at least!), since it implies an ongoing liability to a customer 
who is no longer paying money regularly.  I've wondered if non-expiring in 
combination with a minimum monthly usage would sell, so that eventually 
over at least years, the balance would drain.

The idea that someone could buy a 10 gig block in 2008 and still be 
occasionally using it in 2013, for example, seems odd to me.

>I've read that even ISP's are considering usage-based accounts (to
>penalize you binaries junkies who download 24/7 from unthrottled
>NSP's ;-)

Some ISP's are rumored to be heading in that direction, no doubt.

ISP's frequently buy from backbones using a 95th percentile method of 
billing, so bringing usage billing down to the consumer may be tempting 
for some business models.

Chris Caputo
President, Altopia Corporation

In message <1n6krq.1dt.19.1@news.alt.net> Chris Caputo <ccaputo@alt.net>
wrote:

>In general, with TCP performance testing, I find that a single unthrottled 
>stream is best able to utilize available bandwidth.  Others may disagree 
>or have more scientific input.

For performance testing under ideal circumstances, you're absolutely
correct.

When you're dealing with protocols that cannot be pipelined, add one
additional connection so that between the time one command finishes and
the next can be started, the alternate connection is still downloading.

However, consider that you need the remote side to ACK packets before
the sender can send more data -- As a result, your window size and
latency will provide an upper maximum for the speed of a single TCP
connection, regardless of available bandwidth -- In this situation, you
either use something other then TCP, or open multiple TCP sessions.

To see the effects of latency,
http://www.fancygaphtrn.com/component/option,com_bandwidthcalc/Itemid,5/
has an interesting little calculator.

Enter your data volume and time period, it will tell you the speed
required.  Next, enter your latency and TCP window size and it will
calculate the number of TCP sessions required to fill your bandwidth.

Chris Caputo wrote: 

>thorn <thorn@usenet.> writes:
>>Chris Caputo wrote:
>>>thorn <thorn@usenet.> writes:
>>>>Chris Caputo wrote:
>>>>>thorn <thorn@usenet.> writes:
>>>>>>Chris Caputo wrote:
>>>>>>>Non-customers: Would it increase your chances of joining us?
>>>>>>
>>>>>>Never in a million years, chris.
>>>>>>Prices for a subscription are ridiculous and unattached from reality.
>>>>>
>>>>>How so?  $6/month is ridiculous?
>>>>
>>>>128,000bps/ 8.8 days of retention is a complete joke for six bucks.
>>>>Bet the number of subscribers to this plan can be counted on your
>>>>twelve toes.
>>>>>
>>>>>>Feel free to get acquainted with the plans that other services are
>>>>>>offering and join the rest of us, on planet Earth in the 21st Century.
>>>>>
>>>>>Unless you are getting a better deal than what is offered on your 
>>>>>provider's web site, it looks like our offerings are less expensive.
>>>>
>>>>Au Contraire, oh peanut counter.
>>>>
>>>>Access is granted to the unwashed masses for a paltry sum of ~13.00 -
>>>>20.00 a month from numerous sources. And through the Grace Of God,
>>>>manage to deliver a steady newsfeed that saturates a 15meg connection
>>>>whenever the user decides to logon after paying such an idiotic fee.
>>>>
>>>>Then there are prices for those who just fell off a banana boat:
>>>>    128,000 bits per second max - US$6.00/month
>>>>--
>>>>    2,048,000 bits per second max - US$96.00/month
>>>>
>>>>Which equates to ~720.00 a month to receive the same level of service
>>>>that is delivered for 15.00 from this provider.
>>>
>>>I sincerely appreciate the feedback.
>>>
>>>Taking into account our lower retention, if we offered $12/month 
>>>unthrottled, with SSL and 10 streams, would that be reasonable?
>>
>>			It's an Epiphany! 
>>
>>From: ccaputo@alt.net (Chris Caputo)
>>Subject: Re: Altopia and this group (was: Easynews will filter
>>misplaced binaries!)
>>Date: 8 Aug 2002 22:32:37 GMT
>>Message-ID: <aiuri5$kpc$0@207.14.113.2>
>>
>>thorn <thorn@use.net> writes:
>>>Chris Caputo wrote:
>>>>So all you ex-Altopia customers out there, please come home!!!  Things are
>>>>better and getting even more so. 
>>>
>>>	I am more than willing to "Come Back Home" but your rates are
>>>ridiculous for subscribers with a cable connection. Do you have any
>>>plans to make an adjustment of the subscription fees? 
>>
>>Yes.  In the coming months we will offer unthrottled usage based
>>accounts.
>>
>>Chris Caputo
>>President, Altopia Corporation
>>>
>>>Thanks,
>>>Chris Caputo
>>>President, Altopia Corporation
>
>My how time flies!  :-)
>
>So, tis good?

It's a first step in the right direction - 

Good luck.
>
>Chris Caputo
>President, Altopia Corporation

In article <1n19p7.9e8.19.1@news.alt.net>, ccaputo@alt.net says...
>
>
>thorn <thorn@usenet.> writes:
>>Chris Caputo wrote:
>>>Non-customers: Would it increase your chances of joining us?
>>
>>Never in a million years, chris.
>>Prices for a subscription are ridiculous and unattached from reality.
>
>How so?  $6/month is ridiculous?
>
>>Feel free to get acquainted with the plans that other services are
>>offering and join the rest of us, on planet Earth in the 21st Century.
>
>Unless you are getting a better deal than what is offered on your 
>provider's web site, it looks like our offerings are less expensive.
>
>Chris Caputo
>President, Altopia Corporation
>
Hi Chris,

A few years ago I was one of your subscribers.  I still admire your 
devotion to privacy.

I've been with Newshosting for several years now.  Look at their web 
pages and you may see why.  If you could match their plans, including 
excellent completeness and their long (80-110 days) retention, I'd 
consider returning to you.  Note their secure accelerator.  They also 
have a web to news ('webnews') interface that allows me to check headers 
and get nzb files over a dialup connection.  I don't see webnews 
mentioned on their site right now, but it does exist.

On Thu, 13 Mar 2008 01:17:12 -0000, Chris Caputo <ccaputo@alt.net> wrote:

> engineer2 <engineer2@the.lab> writes:
>> On Wed, 12 Mar 2008 05:25:40 +0000, Chris Caputo wrote:
>>> Taking into account our lower retention, if we offered $12/month
>>> unthrottled, with SSL and 10 streams, would that be reasonable?
>>
>> I've noticed that the number of simultaneous connections (streams?)
>> seems to be important to people, but I've never understood the reason.
>
> I imagine the reasons vary.
>
> With multiple connections you can download binaries in one group while
> browsing text messages in another, for example.
>
>> Does each one of your simultaneous connections add another increment
>> of throttled bandwidth?  Other than that, I can't see any advantage.
>
> On our throttled connections, although our terms indicate that bandwidth
> is distributed across parallel connections, I haven't found it necessary
> to actually implement that and enforce it.  Therefore, on our throttled
> offerings, multiple connections will actually increase overall throughput.
>
> In general, with TCP performance testing, I find that a single unthrottled
> stream is best able to utilize available bandwidth.  Others may disagree
> or have more scientific input.

My newsreader downloads using four connections.  I usually get exactly the bandwidth I've paid for, not four times that.  Occasionally at non-peak times it goes a lot higher, but I assume it's not being limited at those times.

>> I've read that even ISP's are considering usage-based accounts (to
>> penalize you binaries junkies who download 24/7 from unthrottled
>> NSP's ;-)
>
> Some ISP's are rumored to be heading in that direction, no doubt.
>
> ISP's frequently buy from backbones using a 95th percentile method of
> billing, so bringing usage billing down to the consumer may be tempting
> for some business models.

My ISP has an x GB per month for y amount of cash, plus can choose to pay more per GB if you go over the limit, or have it just stop working.

-- 
This message has been brought to you by nuclear power.  Please don't tell the neighbours.
http://www.petersparrots.com    http://www.insanevideoclips.com    http://www.petersphotos.com

              (\
               \)
          ,____/')
          ( __ /
          (/  \)

Peter Hucker wrote:
> On Thu, 13 Mar 2008 01:17:12 -0000, Chris Caputo <ccaputo@alt.net> wrote:
> 
>> engineer2 <engineer2@the.lab> writes:
>>> On Wed, 12 Mar 2008 05:25:40 +0000, Chris Caputo wrote:
>>>> Taking into account our lower retention, if we offered $12/month 
>>>> unthrottled, with SSL and 10 streams, would that be reasonable?
>>> 
>>> I've noticed that the number of simultaneous connections (streams?) 
>>> seems to be important to people, but I've never understood the reason.
>> 
>> I imagine the reasons vary.
>> 
>> With multiple connections you can download binaries in one group while 
>> browsing text messages in another, for example.
>> 
>>> Does each one of your simultaneous connections add another increment of
>>> throttled bandwidth?  Other than that, I can't see any advantage.
>> 
>> On our throttled connections, although our terms indicate that bandwidth 
>> is distributed across parallel connections, I haven't found it necessary 
>> to actually implement that and enforce it.  Therefore, on our throttled 
>> offerings, multiple connections will actually increase overall
>> throughput.
>> 
>> In general, with TCP performance testing, I find that a single
>> unthrottled stream is best able to utilize available bandwidth.  Others
>> may disagree or have more scientific input.
> 
> My newsreader downloads using four connections.  I usually get exactly the
> bandwidth I've paid for, not four times that.  Occasionally at non-peak
> times it goes a lot higher, but I assume it's not being limited at those
> times.

That's a reasonably safe assumption, though it could be buffer under-run
compensation.

> 
>>> I've read that even ISP's are considering usage-based accounts (to 
>>> penalize you binaries junkies who download 24/7 from unthrottled NSP's
>>> ;-)
>> 
>> Some ISP's are rumored to be heading in that direction, no doubt.
>> 
>> ISP's frequently buy from backbones using a 95th percentile method of 
>> billing, so bringing usage billing down to the consumer may be tempting 
>> for some business models.
> 
> My ISP has an x GB per month for y amount of cash, plus can choose to pay
> more per GB if you go over the limit, or have it just stop working.

Sweet!
-- 
ah

On Thu, 20 Mar 2008 10:15:18 -0000, ah <splifingate@gmail.com> wrote:

> Peter Hucker wrote:
>> On Thu, 13 Mar 2008 01:17:12 -0000, Chris Caputo <ccaputo@alt.net> wrote:
>>
>>> engineer2 <engineer2@the.lab> writes:
>>>> On Wed, 12 Mar 2008 05:25:40 +0000, Chris Caputo wrote:
>>>>> Taking into account our lower retention, if we offered $12/month
>>>>> unthrottled, with SSL and 10 streams, would that be reasonable?
>>>>
>>>> I've noticed that the number of simultaneous connections (streams?)
>>>> seems to be important to people, but I've never understood the reason.
>>>
>>> I imagine the reasons vary.
>>>
>>> With multiple connections you can download binaries in one group while
>>> browsing text messages in another, for example.
>>>
>>>> Does each one of your simultaneous connections add another increment of
>>>> throttled bandwidth?  Other than that, I can't see any advantage.
>>>
>>> On our throttled connections, although our terms indicate that bandwidth
>>> is distributed across parallel connections, I haven't found it necessary
>>> to actually implement that and enforce it.  Therefore, on our throttled
>>> offerings, multiple connections will actually increase overall
>>> throughput.
>>>
>>> In general, with TCP performance testing, I find that a single
>>> unthrottled stream is best able to utilize available bandwidth.  Others
>>> may disagree or have more scientific input.
>>
>> My newsreader downloads using four connections.  I usually get exactly the
>> bandwidth I've paid for, not four times that.  Occasionally at non-peak
>> times it goes a lot higher, but I assume it's not being limited at those
>> times.
>
> That's a reasonably safe assumption, though it could be buffer under-run
> compensation.

It's My ISP doing traffic shaping.  Avoid Plusnet.

>>>> I've read that even ISP's are considering usage-based accounts (to
>>>> penalize you binaries junkies who download 24/7 from unthrottled NSP's
>>>> ;-)
>>>
>>> Some ISP's are rumored to be heading in that direction, no doubt.
>>>
>>> ISP's frequently buy from backbones using a 95th percentile method of
>>> billing, so bringing usage billing down to the consumer may be tempting
>>> for some business models.
>>
>> My ISP has an x GB per month for y amount of cash, plus can choose to pay
>> more per GB if you go over the limit, or have it just stop working.
>
> Sweet!

Except for the usenet speed.  Although I could pay more for the gold band.....

-- 
This message has been brought to you by nuclear power.  Please don't tell the neighbours.
http://www.petersparrots.com    http://www.insanevideoclips.com    http://www.petersphotos.com

Confuscious say:  "People who make Confuscious joke speak bad English."

Peter Hucker wrote:
> On Thu, 20 Mar 2008 10:15:18 -0000, ah <splifingate@gmail.com> wrote:
> 
>> Peter Hucker wrote:
>>> On Thu, 13 Mar 2008 01:17:12 -0000, Chris Caputo <ccaputo@alt.net> wrote:
>>>
>>>> engineer2 <engineer2@the.lab> writes:
>>>>> On Wed, 12 Mar 2008 05:25:40 +0000, Chris Caputo wrote:
>>>>>> Taking into account our lower retention, if we offered $12/month
>>>>>> unthrottled, with SSL and 10 streams, would that be reasonable?
>>>>>
>>>>> I've noticed that the number of simultaneous connections (streams?)
>>>>> seems to be important to people, but I've never understood the reason.
>>>>
>>>> I imagine the reasons vary.
>>>>
>>>> With multiple connections you can download binaries in one group while
>>>> browsing text messages in another, for example.
>>>>
>>>>> Does each one of your simultaneous connections add another increment of
>>>>> throttled bandwidth?  Other than that, I can't see any advantage.
>>>>
>>>> On our throttled connections, although our terms indicate that bandwidth
>>>> is distributed across parallel connections, I haven't found it necessary
>>>> to actually implement that and enforce it.  Therefore, on our throttled
>>>> offerings, multiple connections will actually increase overall
>>>> throughput.
>>>>
>>>> In general, with TCP performance testing, I find that a single
>>>> unthrottled stream is best able to utilize available bandwidth.  Others
>>>> may disagree or have more scientific input.
>>>
>>> My newsreader downloads using four connections.  I usually get exactly the
>>> bandwidth I've paid for, not four times that.  Occasionally at non-peak
>>> times it goes a lot higher, but I assume it's not being limited at those
>>> times.
>>
>> That's a reasonably safe assumption, though it could be buffer under-run
>> compensation.
> 
> It's My ISP doing traffic shaping.  Avoid Plusnet.

I did some traffic shaping, once.  Avoid linux.

> 
>>>>> I've read that even ISP's are considering usage-based accounts (to
>>>>> penalize you binaries junkies who download 24/7 from unthrottled NSP's
>>>>> ;-)
>>>>
>>>> Some ISP's are rumored to be heading in that direction, no doubt.
>>>>
>>>> ISP's frequently buy from backbones using a 95th percentile method of
>>>> billing, so bringing usage billing down to the consumer may be tempting
>>>> for some business models.
>>>
>>> My ISP has an x GB per month for y amount of cash, plus can choose to pay
>>> more per GB if you go over the limit, or have it just stop working.
>>
>> Sweet!
> 
> Except for the usenet speed.  Although I could pay more for the gold band.....

Is that GSM, or LTE?
-- 
ah

On Sat, 05 Apr 2008 10:16:12 +0100, ah <splifingate@gmail.com> wrote:

> Peter Hucker wrote:
>> On Thu, 20 Mar 2008 10:15:18 -0000, ah <splifingate@gmail.com> wrote:
>>
>>> Peter Hucker wrote:
>>>> On Thu, 13 Mar 2008 01:17:12 -0000, Chris Caputo <ccaputo@alt.net> wrote:
>>>>
>>>>> engineer2 <engineer2@the.lab> writes:
>>>>>
>>>>> I imagine the reasons vary.
>>>>>
>>>>> With multiple connections you can download binaries in one group while
>>>>> browsing text messages in another, for example.
>>>>>
>>>>>
>>>>> On our throttled connections, although our terms indicate that bandwidth
>>>>> is distributed across parallel connections, I haven't found it necessary
>>>>> to actually implement that and enforce it.  Therefore, on our throttled
>>>>> offerings, multiple connections will actually increase overall
>>>>> throughput.
>>>>>
>>>>> In general, with TCP performance testing, I find that a single
>>>>> unthrottled stream is best able to utilize available bandwidth.  Others
>>>>> may disagree or have more scientific input.
>>>>
>>>> My newsreader downloads using four connections.  I usually get exactly the
>>>> bandwidth I've paid for, not four times that.  Occasionally at non-peak
>>>> times it goes a lot higher, but I assume it's not being limited at those
>>>> times.
>>>
>>> That's a reasonably safe assumption, though it could be buffer under-run
>>> compensation.
>>
>> It's My ISP doing traffic shaping.  Avoid Plusnet.
>
> I did some traffic shaping, once.  Avoid linux.

I always avoid linux.

>>>>> Some ISP's are rumored to be heading in that direction, no doubt.
>>>>>
>>>>> ISP's frequently buy from backbones using a 95th percentile method of
>>>>> billing, so bringing usage billing down to the consumer may be tempting
>>>>> for some business models.
>>>>
>>>> My ISP has an x GB per month for y amount of cash, plus can choose to pay
>>>> more per GB if you go over the limit, or have it just stop working.
>>>
>>> Sweet!
>>
>> Except for the usenet speed.  Although I could pay more for the gold band.....
>
> Is that GSM, or LTE?

No idea what you're talking about, they just give more priority to usenet and p2p.

-- 
This message has been brought to you by nuclear power.  Please don't tell the neighbours.
http://www.petersparrots.com    http://www.insanevideoclips.com    http://www.petersphotos.com

Someday we'll look back on all this and plough into a parked car.

Peter Hucker wrote:
> On Sat, 05 Apr 2008 10:16:12 +0100, ah <splifingate@gmail.com> wrote:
> 
>> Peter Hucker wrote:
>>> On Thu, 20 Mar 2008 10:15:18 -0000, ah <splifingate@gmail.com> wrote:
>>>
>>>> Peter Hucker wrote:
>>>>> On Thu, 13 Mar 2008 01:17:12 -0000, Chris Caputo <ccaputo@alt.net> wrote:
>>>>>
>>>>>> engineer2 <engineer2@the.lab> writes:
>>>>>>
>>>>>> I imagine the reasons vary.
>>>>>>
>>>>>> With multiple connections you can download binaries in one group while
>>>>>> browsing text messages in another, for example.
>>>>>>
>>>>>>
>>>>>> On our throttled connections, although our terms indicate that bandwidth
>>>>>> is distributed across parallel connections, I haven't found it necessary
>>>>>> to actually implement that and enforce it.  Therefore, on our throttled
>>>>>> offerings, multiple connections will actually increase overall
>>>>>> throughput.
>>>>>>
>>>>>> In general, with TCP performance testing, I find that a single
>>>>>> unthrottled stream is best able to utilize available bandwidth.  Others
>>>>>> may disagree or have more scientific input.
>>>>>
>>>>> My newsreader downloads using four connections.  I usually get exactly the
>>>>> bandwidth I've paid for, not four times that.  Occasionally at non-peak
>>>>> times it goes a lot higher, but I assume it's not being limited at those
>>>>> times.
>>>>
>>>> That's a reasonably safe assumption, though it could be buffer under-run
>>>> compensation.
>>>
>>> It's My ISP doing traffic shaping.  Avoid Plusnet.
>>
>> I did some traffic shaping, once.  Avoid linux.
> 
> I always avoid linux.

Why?

> 
>>>>>> Some ISP's are rumored to be heading in that direction, no doubt.
>>>>>>
>>>>>> ISP's frequently buy from backbones using a 95th percentile method of
>>>>>> billing, so bringing usage billing down to the consumer may be tempting
>>>>>> for some business models.
>>>>>
>>>>> My ISP has an x GB per month for y amount of cash, plus can choose to pay
>>>>> more per GB if you go over the limit, or have it just stop working.
>>>>
>>>> Sweet!
>>>
>>> Except for the usenet speed.  Although I could pay more for the gold band.....
>>
>> Is that GSM, or LTE?
> 
> No idea what you're talking about, they just give more priority to usenet and p2p.

Wireless technologies.
-- 
ah

On Thu, 10 Apr 2008 23:50:02 +0100, ah <splifingate@gmail.com> wrote:

> Peter Hucker wrote:
>> On Sat, 05 Apr 2008 10:16:12 +0100, ah <splifingate@gmail.com> wrote:
>>
>>> Peter Hucker wrote:
>>>> On Thu, 20 Mar 2008 10:15:18 -0000, ah <splifingate@gmail.com> wrote:
>>>>
>>>>> Peter Hucker wrote:
>>>>>
>>>>> That's a reasonably safe assumption, though it could be buffer under-run
>>>>> compensation.
>>>>
>>>> It's My ISP doing traffic shaping.  Avoid Plusnet.
>>>
>>> I did some traffic shaping, once.  Avoid linux.
>>
>> I always avoid linux.
>
> Why?

Cheapskate's OS.

>>>>> Sweet!
>>>>
>>>> Except for the usenet speed.  Although I could pay more for the gold band.....
>>>
>>> Is that GSM, or LTE?
>>
>> No idea what you're talking about, they just give more priority to usenet and p2p.
>
> Wireless technologies.

Mobile phones or WIFI?

-- 
This message has been brought to you by nuclear power.  Please don't tell the neighbours.
http://www.petersparrots.com    http://www.insanevideoclips.com    http://www.petersphotos.com

                                  _ .--.
                                 ( `    )
                              .-'      `--,
                   _..----.. (             )`-.
                 .'_|` _|` _|(  .__,           )
                /_|  _|  _|  _(        (_,  .-'
               ;|  _|  _|  _|  '-'__,--'`--'
               | _|  _|  _|  _| |
           _   ||  _|  _|  _|  _|
         _( `--.\_|  _|  _|  _|/
      .-'       )--,|  _|  _|.`
     (__, (_      ) )_|  _| /
       `-.__.\ _,--'\|__|__/
                     ;____;
                      \YT/
                       ||
                      |""|
                      '=='