Mailing List Archive

Hello,

Gopi K Garge wrote:
> IMHO, in the context of the AP region, I guess we are overlooking the
> user perspective - locally available "Internet" content. Barring
> .au, .sg, .kr, .hk (?), .jp and .tw, I do not see any major effort
> around to provide locally, content that is available in the US
> be it the GNU archive, the WU archive or one-of-those Beatles
> song and lyrics archive - result - good amount of traffic to the
> US.

With limited bandwidth, updating these archives would be slow,
and, you wouldn't know yet if there are enough people accessing
these to justify getting a mirror.

Caching should help getting commonly requested information faster.

> When one sees such traffic trends, why would an ISP even think in
> terms of investing in a link to a place other than the US - the cost
> permitting ?

In the long term, as the Internet becomes more common, regional content
will become more important, IMHO. So develop the local content -
and the local infrastructure - and regional intranets, and virtual
private networks - will come in.

--
miguel a.l. paraz <map@iphil.net> iphil communications, makati city,
tech problems, to <support@iphil.net> philippines.





- - - - - - - - - - - - - - - - -

On Thu, 16 May 1996, Gopi K Garge wrote:

> business - NO GO ! Each time we have a meeting, there's a raging
> war on this point, but, bring out the top 25 hosts contacted and
> the related traffic percentages and the consensus is to add more
> bandwidth on the link to the US !

How do you determine the location of those top 25 hosts? If you are
assuming that they are in the US because they are .com or .net, then you
may be wrong.

Michael Dillon Voice: +1-604-546-8022
Memra Software Inc. Fax: +1-604-546-3049
http://www.memra.com E-mail: michael@memra.com

- - - - - - - - - - - - - - - - -

Hola,

] > IMHO, in the context of the AP region, I guess we are overlooking the
] > user perspective - locally available "Internet" content. Barring
] > .au, .sg, .kr, .hk (?), .jp and .tw, I do not see any major effort
] > around to provide locally, content that is available in the US
] > be it the GNU archive, the WU archive or one-of-those Beatles
] > song and lyrics archive - result - good amount of traffic to the
] > US.
]
] With limited bandwidth, updating these archives would be slow,
] and, you wouldn't know yet if there are enough people accessing
] these to justify getting a mirror.

Bah. Transferring it once off-shore is a no-brainer. Of _course_
there will be enough traffic to justify the mirror.

] Caching should help getting commonly requested information faster.

Agreed, hence the policies of Israel and [is it...] New Guinea.
Don't allow traffic over port 80 or 20/21. Make the people use
the cache server.

] > When one sees such traffic trends, why would an ISP even think in
] > terms of investing in a link to a place other than the US - the cost
] > permitting ?
]
] In the long term, as the Internet becomes more common, regional content
] will become more important, IMHO. So develop the local content -
] and the local infrastructure - and regional intranets, and virtual
] private networks - will come in.

Sure, it will become _relatively_ more important than it is now,
but the "Internet" will make geographic location less and less
relevant, and as you allude to, when the bandwidth levels are
increased, the issue of obtaining it locally or from the states,
or from the Moon, shouldn't matter too much.

Yes, I do believe the bw will increase, I don't believe the next
resource (after IP space and IP routes) for us to run out of is
bandwidth. Sure, there's a crunch now, but we'll all just raise
our rates and buy bigger pipes.

Or something.

-alan
- - - - - - - - - - - - - - - - -

Hello,


IMHO, in the context of the AP region, I guess we are overlooking the
user perspective - locally available "Internet" content. Barring
.au, .sg, .kr, .hk (?), .jp and .tw, I do not see any major effort
around to provide locally, content that is available in the US
be it the GNU archive, the WU archive or one-of-those Beatles
song and lyrics archive - result - good amount of traffic to the
US. (The AP region has only 4 archie servers !) When one sees
such traffic trends, why would an ISP even think in terms of
investing in a link to a place other than the US - the cost
permitting ?

I have been attempting to convince ERNET to invest in a link
each to NL/GB and SG/JP since we have a significant amount of
traffic to these places. I have even tried to present the idea
that IN is strategically placed in the context of linking the
Southern AP region to Europe/US which could mean good gatewaying
business - NO GO ! Each time we have a meeting, there's a raging
war on this point, but, bring out the top 25 hosts contacted and
the related traffic percentages and the consensus is to add more
bandwidth on the link to the US !

If on the one hand the tariffs and the telecom policies do not
provide for a good intra-AP connectivity, there is this "content
availability" that's adding to the "US-centricity" (not that
most of the Internet tools aren't ... ;-) ). So ... could we
do anything at all ?

It may not be possible to persuade governments to change their
policies as fast as we would like them to change.

My gut feel is that the tariffs permitting, making content available
locally will hold the key to bandwidth investments and the choice
of the upstream ISP.

--Gopi Garge

Enzo Michelangeli sez:

> On Thu, 16 May 1996, Geoff Huston wrote:

> > Paul,
> >
> > You can be a vocal as you desire, but ultimately from this part of the
> > globe the dominant factor in any ISP business is the cost of the
> > International Private Lease. This lease cost is approximately 10 times
> > the cost of domestic infrastructure.
> >
> > Now when you construct an IPL in a competitive environment where do
> > you terminate it? Generally you are loking for an optimal mix of price
> > and functionality. The observation for the AP region today is that the
> > cheapest IPL half circuits for the AP region terminate in the
> > US. Hence Randy's observation. The internal infrastructure within the
> > AP region happens in a second pass, once the primary objective of
> > major connectivity is achieved internal infrastructure can be cost
> > effective if there is internal traffic flow to match.
> >
> > About the only thing that could hasten regional infrastructure is a
> > drastic revision of the trading practices and expectation of return on
> > investment by the undersea cable investors. Exchange points have
> > little impact per se as they are, in economic terms, a minor aspect of
> > the entire equation.

> Exactly. And all the regional governments should realize that the best way
> of shifting traffic from North America to their region is de-regulating
> the international telecommunications market, scrapping monopolies and
> increasing competition among carriers. That will result much more
> effective than policy-making and verbal "declarations of independence".

> Enzo

--
- - - - - - - - - - - - - - - - -

Hello,

Alan Hannan wrote:

> Yes, I do believe the bw will increase, I don't believe the next
> resource (after IP space and IP routes) for us to run out of is
> bandwidth. Sure, there's a crunch now, but we'll all just raise
> our rates and buy bigger pipes.

Bandwidth increases in both within the US and in bandwidth-starved
countries, but will the price ratio go down?

Currently, the $$$ our ISP pays to get 128 Kbps connectivity is
enough to get more than a dozen T1's in the US, and even a T3
perhaps. And people ask here why they can't get unlimited access for
$20 monthly.

On the other hand, labor costs are cheap --- among the
biggest customers for E1 lines to the US are offshore programming
houses.

So how about a bandwidth for workforce swap? :)

--
miguel a.l. paraz <map@iphil.net> iphil communications, makati city,
tech problems, to <support@iphil.net> philippines.





- - - - - - - - - - - - - - - - -

> Currently, the $$$ our ISP pays to get 128 Kbps connectivity is
> enough to get more than a dozen T1's in the US, and even a T3
> perhaps. And people ask here why they can't get unlimited access for
> $20 monthly.
>
> On the other hand, labor costs are cheap --- among the
> biggest customers for E1 lines to the US are offshore programming
> houses.
>
> So how about a bandwidth for workforce swap? :)
>
> --
> miguel a.l. paraz <map@iphil.net> iphil communications, makati city,
> tech problems, to <support@iphil.net> philippines.
>

Been There, Done That.

15 years ago, the company I worked for dropped a 50meter earth
station and the rights to four transponders for 25 years on a
project that was donated to a non-us telco to reduce the costs
of transporting and housing 300 engineeers a quarter to/from
the US. It was a very cheap alternative.

--
--bill
- - - - - - - - - - - - - - - - -

At 06:57 AM 5/16/96 +1000, Geoff Huston wrote:
>Paul,
>
>You can be a vocal as you desire, but ultimately from this part of the
>globe the dominant factor in any ISP business is the cost of the
>International Private Lease. This lease cost is approximately 10 times
>the cost of domestic infrastructure.
>
>Now when you construct an IPL in a competitive environment where do
>you terminate it? Generally you are loking for an optimal mix of price
>and functionality. The observation for the AP region today is that the
>cheapest IPL half circuits for the AP region terminate in the
>US. Hence Randy's observation. The internal infrastructure within the
>AP region happens in a second pass, once the primary objective of
>major connectivity is achieved internal infrastructure can be cost
>effective if there is internal traffic flow to match.
>
>About the only thing that could hasten regional infrastructure is a
>drastic revision of the trading practices and expectation of return on
>investment by the undersea cable investors. Exchange points have
>little impact per se as they are, in economic terms, a minor aspect of
>the entire equation.

Yes, they are still a minor aspect but I think they contribute serveral more
percents to the return on investment. In AP, we have exchange points such
as NSPIXP of Japan and HKIX of Hong Kong. If ISPs in other AP countries
have links to JP or HK and automatically (or semi-automatically) they can
connect to all ISPs in that country/region instead of just one or several
ISPs in it, the connection is more valuable and justifiable. If the links
can also be used to connect to other ISPs in other AP countries via that
exchange, the connection is even more valuable and justifiable. Of course,
we still can't eliminate the links to US altogether. But we probably can
slow down the growth of bandwidth requirements to US a little bit.

(NOTE: HKIX has mandatory MLPA for routes within Hong Kong.)

Che-Hoo Cheng | Email: chcheng@cuhk.edu.hk
Data Comms and Networking Section | URL: http://www.cuhk.edu.hk/
Information Technology Service Unit | Tel: +852 2609-8848
The Chinese University of Hong Kong | Fax: +852 2603-5001

- - - - - - - - - - - - - - - - -

Miguel A.L. Paraz sez:

> Hello,

> Gopi K Garge wrote:
> > IMHO, in the context of the AP region, I guess we are overlooking the
> > user perspective - locally available "Internet" content. Barring
> > .au, .sg, .kr, .hk (?), .jp and .tw, I do not see any major effort
> > around to provide locally, content that is available in the US
> > be it the GNU archive, the WU archive or one-of-those Beatles
> > song and lyrics archive - result - good amount of traffic to the
> > US.

> With limited bandwidth, updating these archives would be slow,
> and, you wouldn't know yet if there are enough people accessing
> these to justify getting a mirror.

Oh ... I have complete details on how many hits some of these
servers get from my domain and who in my domain are originating
requests. getting these details is a matter of some dedicated
hardware and a one time programming effort - the rest is -
thanks to ethernet technology.

> Caching should help getting commonly requested information faster.

Sure ... but you'd notice that these are mechanisms to control
your traffic across a low bandwidth link, but that's not the
issue. The issue is where will the link that we have terminate.

> > When one sees such traffic trends, why would an ISP even think in
> > terms of investing in a link to a place other than the US - the cost
> > permitting ?

> In the long term, as the Internet becomes more common, regional content
> will become more important, IMHO. So develop the local content -
> and the local infrastructure - and regional intranets, and virtual
> private networks - will come in.


When ?? I feel it is long overdue. If we are to "attract" intra
AP region bandwidth investment - as in getting ISPs to interconnect
within the region, there must be an effort to provide
content. The ISPs may then have a case (not a strong one though)
to interconnect locally. How do we do this ? I know that this is
not a trivial question, but ......

--Gopi
- - - - - - - - - - - - - - - - -

At 12:44 AM 5/18/96 +0800, Che-Hoo Cheng wrote:

>
>Yes, they are still a minor aspect but I think they contribute serveral more
>percents to the return on investment. In AP, we have exchange points such
>as NSPIXP of Japan and HKIX of Hong Kong. If ISPs in other AP countries
>have links to JP or HK and automatically (or semi-automatically) they can
>connect to all ISPs in that country/region instead of just one or several
>ISPs in it, the connection is more valuable and justifiable. If the links
>can also be used to connect to other ISPs in other AP countries via that
>exchange, the connection is even more valuable and justifiable. Of course,
>we still can't eliminate the links to US altogether. But we probably can
>slow down the growth of bandwidth requirements to US a little bit.
>

Thanks for mentioning this -- it is exactly what is needed. :-)

- paul

- - - - - - - - - - - - - - - - -

In article <QQapom01899.199605132105@rodan.UU.NET>,
Mike O'Dell <mo@uunet.uu.net> wrote:
>one other solution (being implemented, I believe), is a DNS server
>that listens to the BGP traffic, so it knows how far away things are,

Another option is to take advantage of the timing information DNS already
keeps track of:

$ORIGIN foobar.com.
www IN NS www1
www IN NS www2
www IN NS www3

Each of those nameservers is actually the webserver... and responds only
with the A record for itself. Turn down the TTL. Let DNS caching
take care of the rest...

It doesn't work because, as far as I've seen, the clients (ok, netscape)
doesn't ever requery the DNS. I'm not sure what state they're at now --
but they're moving towards doing all the resolver work themselves so as
to get multithreaded behaviour. Maybe we can convince the browser writers
to use ttl.

Dean
- - - - - - - - - - - - - - - - -