🚀 go-pugleaf

...so anyway, this dude wingo@asiaonline.net (Wingo Luk) said:

>__R_e_t_r_o__ (retro@asiaonline.net) wrote:

>: I have virtually given up using the Web (with Asia On-Line). Firstly, if the DNS
>: ...[snip]...

>	The Speed in WAN is definitly slow than in LAN...you can't 
>request a internet connection with a speed of your modem all the time.
>	Also .the connection to US is not as simple as connection of Two 28800 
>modem within HK. Once the connection is established, the line have to face 
>a lot of problem like congestion or overload traffic problem. It doesn't 
>like a modem connection, you can get use of the whole bandwidth...

...oh, I see, do you suppose this is why I said in my original post "assuming
that remote host response is *perfect*"? Thus allowing this simplistic model to
have some relevance?

>: Of course, I wasn't even able to connect to HKStar (who apparently do not have a
>: T1 HKIX connection) to get the phone number to apply for a new Internet account.
>	HkStar have a T1 to HKIX , you can't connect to Star, why you 
>think that the Stax is downed instead of AOL get problems....

Well, as you know Wingo, Star, it seems, do not have a T1 to HKIX, meaning that
in all likeliness, access to them from AOL would be routed via the 256Kbps to
the US.

>: Oh, so they are charging for changes now - many other ISP's in the world don't
>: even charge a set up fee for users having web pages, since it is perfectly
>: feasible for users to service pages themselves.
>	AOL use seperate Servers for Unix and WWW in order to 
>reduce the loading of each server, therefore, it require human to setup 
>the pages for you. 

Perhaps this is so, but *all* that a user needs is FTP access to the directory
where his/her pages are stored in order to post, maintain and change a web page
*themselves*.

>: ...and now the interesting bit:

>: >                        256Kbps Sprintlink to US Sprint 
>: >                        T1 (1.544Mbps) to USA (November 95)
>:       We can only hope-  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>	Read the news in aol.announce....

Oh, you mean the bit where John Thompson says;

"In response to our users' demands for faster speed, Asia On-Line has
committed to putting a T1 line direct to the US Internet backbone and
are actively negotiating with carriers now.  We will make a definite
announcement on this within November."

Note how he says an 'announcement' will be made in Nov., not that it will be
implemented in Nov.

>: >Over 250 dial up lines
>: >All lines supported by Hayes 28.8Kbps modems
>:                               ^^^^^^^^^^^^^^^  Why?
>	Why? Hayes modem have good quality and can compatible with most 
>of the modem in HK...

Yes, and I was being facetious, asking "Why?". In the context of my original
post, why have 28.8Kbps modems when you can hardly reach speeds worthy of a
9,600bps modem over the 256Kbps link?

>: ...[snip]...

>: Okay, so at present the *only* international link is 256Kbps to the US.

>: - Let's just say 125 people have dialed up, out of 250 lines (50%)
>: - Let's say 80% are connecting to sites via the 256Kbps line
>: - Let's say given that people using the Web have to *view* their chosen page and
>: are not continually trying to transfer data (but bearing in mind that some
>: people will be using FTP or perhaps RealAudio) that each person is trying to
>: transfer data 30% of the time.

>	I assume the data and equation are correct ! but it seems that you have 
>misunderstand between bit and byte.. Check carefully...

small b = bit, big B = Byte, do you see any big B's anywhere? Perhaps I confused
you by changing between Kbps and bps units too often? Once *you* have your units
figured, try it with *your* solar powered calculator.

The data of course is a mixture of actual figures from AOL, and hypothetical
guesstimates.

>: 256 / (125 x 80% x 30%) = 8.533
>  ^^^^	
>: So the fastest data transfer I might expect in this model (assuming that remote
>: host response is *perfect*) is 8,533 bps. Which in fact, is what I *had* become
> ...[snip]...

>: Speaking of money, let's just say out of the total user connect capacity of AOL,
>: they get 30% utilisation (Again a rough model, ...[snip]...

>: but how long will it be before AOL charges a *realistic* connect charge that can
>: support growth *and* is able to provider users with an 'effective' data transfer
>: speed?
>	Why don't quote some figure to prove your argument above....

I did say "Again a rough model"

I am not trying to prove anything, just trying to stimulate some thought on the
subject  :]

Oops, am I advocating *higher* connect charges? <gasp> Actually, I would much
rather see the extra amount being paid for PNETS going into the pockets of the
ISPs to support growth than into the pockets of some government authority that
doesn't even know what they are dealing with!

>--
>                <00>
>===============H====H============================================
>Wingo Luk Wai Kwok 
>City University of Hong Kong      BEng (Hons) Computer Engineering 
>E-mail : 93846370@cityu.edu.hk , wingo@asiaonline.net 
>	 wingo.luk@hknet.org  , wingo@PcSoftlink.org.hk
>URL : http://www.PcSoftlink.org.hk/~wingo
>==============/=====\============================================ 

Since I like making 'guesses', I would guess that:

a) Your web-surfing via your AOL account is giving you value for money out of
having a V34 modem?

or,
b) You have a better explanation why data-transfer via AOL has virtually come to
a halt?

or,
c) You do your web-surfing at CUHK during your study breaks?


___Richard de Nys___

retro@asiaonline.net