I just don't understand what the problem is. WHY do people wish to reduce
the pop connector collection time?

Well, actually, NO, that's not the question. The question is more like 'WHY
do people put so much effort into reducing the pop connector retrieval
interval?'. With maybe a corollary question of 'WHY is it so damned
important?'.

I was dealing with a company for tech support today, the guy emailed me some
notes. As we receive via SMTP no POP connector delay was involved. While I
was on the phone I heard him reading the email, he suggested he was
forwarding it to me and I could hear him typing as he asked me to pause
while I was giving him my email address, he then suggested 'you should get
it soon' so I'm supposing he had actually hit the 'send' button at this
time.

I also called someone else for support yesterday and went through a similar
scenario.

Both emails were simple text messages. One arrived inside the 15 minutes,
the other took 3 hours. One was from Trend Micro, the other was from
Microsoft. This is the nature of internet email. An old quote about internet
email suggests '90% gets delivered inside a few minutes, 5% takes a couple
of days, 5% goes into the big pop box in the sky and no-one knows why.'

I also receive via SMTP at home. If the girl next door emailed me I would
receive it with minimal delay. If she expected me to get off my ass and
knock on her door within 15 minutes she better be offering something better
than dinner to fix her computer.

I just don't understand what the frenzy is.

On the other hand, there are very good reasons why after attempting POP
collection once a short delay should occur before another pop collection is
attempted. Many mail systems (and I'm probably talking *nix here,
particularly *nix mail server farms) 'lock' the mailbox during delivery, if
you attempt to access the mailbox before the lock is released it will error.
The previous POP conversation can be complete but because the server is
handling thousands of requests and also doing it's own housekeeping the
mailbox may still be locked. I'm not sure but I think the lock exists to
eradicate the chance of duplicate delivery.

ie. (if the mailbox lock was not implemented)
popconversation1 'HI, any new mail?'
popconversation1 'Yep, sure is, 1*10MB message'
popconversation1 'give it to me baby'
Transfer starts
popconversation2 'HI, any new mail?'
popconversation2 'Yep, sure is, 1*10MB message'
popconversation2 'give it to me baby'
Transfer starts

some time later more pop conversations get initiated and we've got more
transfers of the same message which eventually may end. The client then has
multiple copies of the same message which transferred slowly because they
all happened at the same time,,, and gets rather confused what to do with
them.

Just my take on the issue. I must be 'in a mood'. IF your users DEMAND
immediate reception of mail tell them the best you can do is a permanent IP
and SMTP delivery, which may result in delays of several minutes to several
days, but won't include an additional 14 minutes and 59 seconds.

Re: the 15 minute pop equation by Dave

Dave
Tue Jan 13 07:09:06 CST 2004

LOL



Re: the 15 minute pop equation by James

James
Tue Jan 13 07:14:45 CST 2004

SuperGumby wrote:
> I just don't understand what the problem is. WHY do people wish to
> reduce the pop connector collection time?[snip]

...prompting yet another question: why even discuss the 15 minute interval
with end users? If we all just set up SBS boxes and said "it will check
your mail for you" and don't add "every 15 minutes" then nobody would be the
wiser IMHO.

It's not like most SBS servers have multiply-redundant load-balanced T3
internet connections :-) On a 56k modem it really doesn't matter how often
you check your pop3 mailboxes - it takes ages to download the messages...

James



Re: the 15 minute pop equation by NoJags

NoJags
Tue Jan 13 07:44:20 CST 2004


"SuperGumby" <not@your.nellie> wrote in message
news:%23IovdXd2DHA.3436@tk2msftngp13.phx.gbl...
<snip>
>
> Just my take on the issue. I must be 'in a mood'.

Hey, no problem. You want to see me in one of *my* moods.

>IF your users DEMAND
> immediate reception of mail tell them the best you can do is a permanent
IP
> and SMTP delivery, which may result in delays of several minutes to
several
> days, but won't include an additional 14 minutes and 59 seconds.
>

In the case of my company, the ISP originally sold the internet connection
on the basis that email was "more or less instant". I remember the meeting
(c.1996) clearly; I knew better but I kept my mouth shut because I *wanted*
an internet connection. Of course, when it turned out that email wasn't
instant, it fell on me to do something about it so we ended up running our
own mailserver on a fixed IP address, which IMHO is stupid for a company of
our size.



Re: the 15 minute pop equation by Daniel

Daniel
Tue Jan 13 13:21:03 CST 2004

There is also the issue with the additional cost of the fixed Ip address,
here in New Zealand it costs $30 a month for one. Believe it or not alot of
business's don't wish to pay telecom this much extra. I just tell the
client it checks every 15 minutes and they are fine with that.

Just my two cents

Daniel.

"NoJags Neil" <neil264removethisbit@hotmail.com> wrote in message
news:eAvDitd2DHA.484@TK2MSFTNGP10.phx.gbl...
>
> "SuperGumby" <not@your.nellie> wrote in message
> news:%23IovdXd2DHA.3436@tk2msftngp13.phx.gbl...
> <snip>
> >
> > Just my take on the issue. I must be 'in a mood'.
>
> Hey, no problem. You want to see me in one of *my* moods.
>
> >IF your users DEMAND
> > immediate reception of mail tell them the best you can do is a permanent
> IP
> > and SMTP delivery, which may result in delays of several minutes to
> several
> > days, but won't include an additional 14 minutes and 59 seconds.
> >
>
> In the case of my company, the ISP originally sold the internet connection
> on the basis that email was "more or less instant". I remember the
meeting
> (c.1996) clearly; I knew better but I kept my mouth shut because I
*wanted*
> an internet connection. Of course, when it turned out that email wasn't
> instant, it fell on me to do something about it so we ended up running our
> own mailserver on a fixed IP address, which IMHO is stupid for a company
of
> our size.
>
>



Re: the 15 minute pop equation by Andrew

Andrew
Tue Jan 13 19:44:58 CST 2004

My own question these days is "Why is *everything* so important?" It
becomes very difficult to set priorities when you have 30-40 active clients
and every little problem needs to be fixed immediately. Sometimes I feel
like a drug pusher; I just have computer junkies instead of drug junkies.

"SuperGumby" <not@your.nellie> wrote in message
news:%23IovdXd2DHA.3436@tk2msftngp13.phx.gbl...
> I just don't understand what the problem is. WHY do people wish to reduce
> the pop connector collection time?
>
> Well, actually, NO, that's not the question. The question is more like
'WHY
> do people put so much effort into reducing the pop connector retrieval
> interval?'. With maybe a corollary question of 'WHY is it so damned
> important?'.
>
> I was dealing with a company for tech support today, the guy emailed me
some
> notes. As we receive via SMTP no POP connector delay was involved. While I
> was on the phone I heard him reading the email, he suggested he was
> forwarding it to me and I could hear him typing as he asked me to pause
> while I was giving him my email address, he then suggested 'you should get
> it soon' so I'm supposing he had actually hit the 'send' button at this
> time.
>
> I also called someone else for support yesterday and went through a
similar
> scenario.
>
> Both emails were simple text messages. One arrived inside the 15 minutes,
> the other took 3 hours. One was from Trend Micro, the other was from
> Microsoft. This is the nature of internet email. An old quote about
internet
> email suggests '90% gets delivered inside a few minutes, 5% takes a couple
> of days, 5% goes into the big pop box in the sky and no-one knows why.'
>
> I also receive via SMTP at home. If the girl next door emailed me I would
> receive it with minimal delay. If she expected me to get off my ass and
> knock on her door within 15 minutes she better be offering something
better
> than dinner to fix her computer.
>
> I just don't understand what the frenzy is.
>
> On the other hand, there are very good reasons why after attempting POP
> collection once a short delay should occur before another pop collection
is
> attempted. Many mail systems (and I'm probably talking *nix here,
> particularly *nix mail server farms) 'lock' the mailbox during delivery,
if
> you attempt to access the mailbox before the lock is released it will
error.
> The previous POP conversation can be complete but because the server is
> handling thousands of requests and also doing it's own housekeeping the
> mailbox may still be locked. I'm not sure but I think the lock exists to
> eradicate the chance of duplicate delivery.
>
> ie. (if the mailbox lock was not implemented)
> popconversation1 'HI, any new mail?'
> popconversation1 'Yep, sure is, 1*10MB message'
> popconversation1 'give it to me baby'
> Transfer starts
> popconversation2 'HI, any new mail?'
> popconversation2 'Yep, sure is, 1*10MB message'
> popconversation2 'give it to me baby'
> Transfer starts
>
> some time later more pop conversations get initiated and we've got more
> transfers of the same message which eventually may end. The client then
has
> multiple copies of the same message which transferred slowly because they
> all happened at the same time,,, and gets rather confused what to do with
> them.
>
> Just my take on the issue. I must be 'in a mood'. IF your users DEMAND
> immediate reception of mail tell them the best you can do is a permanent
IP
> and SMTP delivery, which may result in delays of several minutes to
several
> days, but won't include an additional 14 minutes and 59 seconds.
>
>



Re: the 15 minute pop equation by Javier

Javier
Tue Jan 13 21:33:45 CST 2004

> There is also the issue with the additional cost of the fixed Ip address,
> here in New Zealand it costs $30 a month for one. Believe it or not alot
of
> business's don't wish to pay telecom this much extra. I just tell the
> client it checks every 15 minutes and they are fine with that.

You don't need a fixed IP to host your own mail... just get a (free) dynamic
DNS service to do the trick (i.e. www.dyndns.org). The only "bad thing"
would be if your ISP is blocking port 25 inbound (which can also be
solved... but this time requires some green).

Cheers!

--
-Javier

<< SBS ROCKS !!! >>



Re: the 15 minute pop equation by Javier

Javier
Tue Jan 13 21:35:23 CST 2004

>> Of course, when it turned out that email wasn't
>> instant, it fell on me to do something about it so we ended up running
our
>> own mailserver on a fixed IP address, which IMHO is stupid for a company
>> of our size.

I dunno about that being stupid...

I have a server on my home hosting my own mail (with only one user + the
dog). Is that so bad??? :-)

--
-Javier

<< SBS ROCKS !!! >>



Re: the 15 minute pop equation by NoJags

NoJags
Wed Jan 14 09:02:06 CST 2004


"Javier Gomez" <javier_gomez@remove.this.bit.engineer.com> wrote in message
news:ed3Oy9k2DHA.3436@tk2msftngp13.phx.gbl...
> >> Of course, when it turned out that email wasn't
> >> instant, it fell on me to do something about it so we ended up running
> our
> >> own mailserver on a fixed IP address, which IMHO is stupid for a
company
> >> of our size.
>
> I dunno about that being stupid...
>
> I have a server on my home hosting my own mail (with only one user + the
> dog). Is that so bad??? :-)
>

If you know how to do that without making a pigs ear of it, and don't mind
spending the time then it's fine. The reason that I believe it was stupid
for a small company like ours was that I had to spend *a lot* of man-hours
learning how to do it when I should have been doing other stuff, made lots
of mistakes along the way, and still have to spend a lot of time
maintaining it. Far better subbed-out imho.



Re: the 15 minute pop equation by Daniel

Daniel
Wed Jan 14 13:16:54 CST 2004

yes I know where you are coming from. I was really annoyed one day...One of
my clients has a 15 computer network. Since I put in the NT4 Server with
Exchange 5.5 they have been running without a hitch for 2 years. One day I
was there and the hub decided to lock up, it took me 10 minutes to get
everything operational again. The way they were carrying on you would have
thought the world was going to end! Gratitude for you!

Daniel.


"Andrew M. Saucci, Jr." <andrew@2000computer.com> wrote in message
news:u2ZqGAk2DHA.2336@TK2MSFTNGP09.phx.gbl...
> My own question these days is "Why is *everything* so important?"
It
> becomes very difficult to set priorities when you have 30-40 active
clients
> and every little problem needs to be fixed immediately. Sometimes I feel
> like a drug pusher; I just have computer junkies instead of drug junkies.
>
> "SuperGumby" <not@your.nellie> wrote in message
> news:%23IovdXd2DHA.3436@tk2msftngp13.phx.gbl...
> > I just don't understand what the problem is. WHY do people wish to
reduce
> > the pop connector collection time?
> >
> > Well, actually, NO, that's not the question. The question is more like
> 'WHY
> > do people put so much effort into reducing the pop connector retrieval
> > interval?'. With maybe a corollary question of 'WHY is it so damned
> > important?'.
> >
> > I was dealing with a company for tech support today, the guy emailed me
> some
> > notes. As we receive via SMTP no POP connector delay was involved. While
I
> > was on the phone I heard him reading the email, he suggested he was
> > forwarding it to me and I could hear him typing as he asked me to pause
> > while I was giving him my email address, he then suggested 'you should
get
> > it soon' so I'm supposing he had actually hit the 'send' button at this
> > time.
> >
> > I also called someone else for support yesterday and went through a
> similar
> > scenario.
> >
> > Both emails were simple text messages. One arrived inside the 15
minutes,
> > the other took 3 hours. One was from Trend Micro, the other was from
> > Microsoft. This is the nature of internet email. An old quote about
> internet
> > email suggests '90% gets delivered inside a few minutes, 5% takes a
couple
> > of days, 5% goes into the big pop box in the sky and no-one knows why.'
> >
> > I also receive via SMTP at home. If the girl next door emailed me I
would
> > receive it with minimal delay. If she expected me to get off my ass and
> > knock on her door within 15 minutes she better be offering something
> better
> > than dinner to fix her computer.
> >
> > I just don't understand what the frenzy is.
> >
> > On the other hand, there are very good reasons why after attempting POP
> > collection once a short delay should occur before another pop collection
> is
> > attempted. Many mail systems (and I'm probably talking *nix here,
> > particularly *nix mail server farms) 'lock' the mailbox during delivery,
> if
> > you attempt to access the mailbox before the lock is released it will
> error.
> > The previous POP conversation can be complete but because the server is
> > handling thousands of requests and also doing it's own housekeeping the
> > mailbox may still be locked. I'm not sure but I think the lock exists to
> > eradicate the chance of duplicate delivery.
> >
> > ie. (if the mailbox lock was not implemented)
> > popconversation1 'HI, any new mail?'
> > popconversation1 'Yep, sure is, 1*10MB message'
> > popconversation1 'give it to me baby'
> > Transfer starts
> > popconversation2 'HI, any new mail?'
> > popconversation2 'Yep, sure is, 1*10MB message'
> > popconversation2 'give it to me baby'
> > Transfer starts
> >
> > some time later more pop conversations get initiated and we've got more
> > transfers of the same message which eventually may end. The client then
> has
> > multiple copies of the same message which transferred slowly because
they
> > all happened at the same time,,, and gets rather confused what to do
with
> > them.
> >
> > Just my take on the issue. I must be 'in a mood'. IF your users DEMAND
> > immediate reception of mail tell them the best you can do is a permanent
> IP
> > and SMTP delivery, which may result in delays of several minutes to
> several
> > days, but won't include an additional 14 minutes and 59 seconds.
> >
> >
>
>