David Garcia on 13 Jan 2001 14:36:00 -0000

[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

<nettime> 13 reactions

The following posting to a closed list for a project called CIRCUS (Content
Integrated Research into User Systems...sigh) on developing strategies for
better communication for people from the arts and the sciences. I post this
here because I introduced the list to Gerry McGovern's 13 Things to know
about Broadband which I first found posted on nettime. It drew this lively
response from Glasgow University's delegate to CIRCUS
From: John W Patterson <jwp@dcs.gla.ac.uk>
To: David Garcia <davidg@xs4all.nl>
Subject: This is the one to post
Date: Wed, Jan 10, 2001, 14:12

David responded to John by suggesting that he look at
"13 Things to Know About Broadband" by Gerry McGovern for a more sceptical
perspective on the broadband future. An extract a book on streaming media
(Net.Congestion) which I am co-editing.

John replied...

Your quotes suggest to me that you are skeptical of my view that
web-mediated multimedia, including streaming video will come to
dominate telematics-based learning and academic-style learning
generally. My response is that, if you look closely at your quotes,
they don't actually point to any show-stoppers in the scenario I am

>13 Things to Know About Broadband
>Gerry McGovern
>1. While fiber optic cable is being rolled out at a frantic pace it is
>barely keeping up with overall bandwidth demand. "In the next five years we
>don't see any ability of service providers in the U.S. to keep up with the
>demand," Mouli Ramani, director of strategic marketing for the optical
>Internet at Nortel Networks told Inter@ctive Week in February 2000. "I don't
>see any chance of getting into a glut anywhere in the network over the next
>five years."

I recall saying something very similar at BKSTS VFX 2000 in
September. They were all talking about how TiVO and its like were
going to make advertising obsolete and the delivery of movies over
the Internet was going to destroy the economics of film distribution.
My comment in essence was 'don't hold your breath', and was based on
the technical observation that, while ADSL offered 2 Megabits
(MB)/sec that was only to the one person on the local loop; if
everyone in the street got connected you got that fraction - if you
were lucky 56K - at ADSL prices (its bog standard 10MB/s ethernet

Its the local loop which is the broadband killer on the public
network. Its the topology that is the killer for broadband cable
services. Both can be rectified.

>2. The backbone of the Internet hasn't been designed to deal with millions
>of people having broadband access.

Media-rich web-pages don't need to use the external backbone. For
example, if you are selling a teaching package you can load your key
resources - including streamable video - on a local server.

>3. The flat-fee 'use as much bandwidth as you like' Internet pricing model
>is unworkable in a broadband environment where one user might want to use
>hundreds or thousands time more bandwidth than their neighbor.

Sure, you'll be paying for your bandwidth but if you're a big
institution you'll have to decide whether to hire your bandwidth or
install it in-house. Many institutions have gone for 'in-house'.
Digital TV is deigned round this model. You can pay for a low-res
feed, the same sort of res as you have now or HDTV-res which is 4-6
times as expensive because it uses 4-6 times the bandwidth.

>4. The Internet works from a weakest link in the chain point of view, so
>just because you have broadband access doesn't mean that a particular
>website will download any faster.

Ditto. I'm sitting on a gigabit ethernet right now, and I can take
streaming video from the USA, typically at 100+KB/sec.

>5. Because of Internet limitations, broadband suppliers are increasingly
>choosing private high-capacity networks to deliver their services to
>subscriber-based audiences.

As for my answer to 3.

>6. As yet, no broadband online entertainment companies have gone public and
>many, such as Digital Entertainment Network, have shed staff and re-focused
>their business models away from creating original broadband content.

..and Pseudo went bust (amid rumours which suggested that their
collapse wasn't wholly due to their business model being wrong), and
what about atom.com and its chums? (Actually its very much alive and
as sassy as ever at http://www.atomfilms.com.) The word is: broadband
is coming, maybe 2 years down the line. The surviving early adopters
and experimenters will then be in a very strong position.

>7. Yahoo and Lycos have scaled back content and service plans for broadband
>users, citing the basic fact that for every broadband user there are 50 with
>basic access.

If the US experience is anything like the UK one its because the
cable providers have been dragging their feet over digital access.
However, in the US cable is dominant, here its struggling to get in
ahead of satellite - a niche market over there. You'd think that this
would make the UK cable industry a bit more nimble than it actually
is but in reality the one organisation which would make them get off
their back-sides, BT, isn't allowed to play (yet). The reason they're
so sluggish is that consolidation has made for regional monopolies in
cable services and, outside London, there are only two cable
companies, both branch factories of US-based companies who don't do
anything unless their US headquarters tells them. God knows what the
UK industry looks like to their bean-counters, probably the
unprofitable bit they're reluctant to throw good money after bad at.

There is the issue that once users get their PCs they're reluctant to
change (unless Microsoft forces them to because their software has
suddenly become incompatible with everyone elses'), just like they're
reluctant to change their TVs (a major drag on the HDTV market). Most
PCs out there have only got the most basic web browser and the most
basic capabilities. Here there is a ray of hope because most PCs are
made so badly that they die after a few years, forcing the user to
buy a machine inevitably hugely more powerful and capable than its
predecessor. When that happens watch demand soar.

>8. Broadband access providers such as Pacific Bell, Midwestern and @Home
>have been suffering severe email service slowdowns recently, as broadband
>users send huge video and sound files as email attachments.

..suggesting there is the demand! This is simply a matter of the
service providers putting in the bandwidth. The story goes that the
UK installed the necessary bandwidth yonks ago and BT is just sitting
on it, waiting for the day they're allowed to provide content.
Meantime their share price takes a bashing because everyone's begun
to realise that the parts are worth more than the sum.

>9. Because broadband generally establishes 24-hour-a-day connections to the
>Internet, it creates a serious security threat. "The home user is more
>susceptible to someone coming and stealing the information that is on their
>computer," David Remnitz, chief executive of IFsec LLC, a New York network
>security firm told Nando Times in February 2000. "They could be monitoring
>messages that are sent in or out of that system, which could be things like
>bank routing numbers."

This is at core a red herring. By far the more serious security
problem is that of denial-of-service attacks and those intended to
overwhelm security defences. This is usually done by co-opting
thousands of inadequately protected machines to bombard the target
with fake IP packets or the like. This happens now. As for the humble
user its like being a single piece of straw in a haystack. You are
far more likely to be burned by the fire caused by the lightning
which hits the stack than be fried by the lightning bolt itself.
Another analogy: how many homeowners have really effective burglar
deterrents, and do those that don't keep their valuables/money at

The only sensible place for the bad guys to monitor for this stuff is
at the ISP end and they (the ISPs) should be investing in adequate
security. A new secure internet IP is coming in which should help
ISPs a lot with this.

A head of an internet security company has a vested interest in
panicking people about home security.

>10. In March, The Wall Street Journal reported that because traffic was
>overwhelming the broadband networks of cable companies, the numbers of
>houses served by a single cable 'node' were being reduced from 10,000, as
>originally projected, to 500 or fewer. Many cable companies were being
>forced to monitor individual usage.

See my answers to 1 and 2. In respect to high-resource, high
production-value telematics- based teaching this is a non-issue.

In any case the above is a demonstration of incompetence. A simple
back-of- the-envelope calculation would have made the above obvious.
If it can occur to me( a non-expert in respect to internet technology
or telecommunications) - see answer 1 - then it can occur to the
likes of managers in cable companies. If you're pumping out the same
signals to everyone you're limited only by the bandwidth of the
communications medium. If everyone decides what their input is going
to be individually then every house has to have bandwidth equal to
its own channel and if you've been used to pumping out the same stuff
you'll have a tree network which means the 'root' of the tree will
have to have as much bandwidth as the sun of the houses it serves.
The answer is not to have a tree but a network with a greater
'redundancy' of channels - people are thinking of a highly
interconnected core (a bit like the telephone system) with
lowly-connected but overlapping strands leading out into the customer ends.

Its a great pity BYTE isn't around any more. This is just the sort of
stuff they could present so well for non-specialists - and did, to
the benefit of all.

>11. The popularity of Napster and other such devices, which allow people to
>easily swap music files, has slowed many university campus networks to a
>crawl. Some universities estimated that Napster downloads were using more
>than 50 percent of their available bandwidth.

Yet I'm told Stanford have managed to put streaming video feeds for their
lectures on their campus for years. I don't know the core technology
but VOD and NVOD experiments were done in the area some years ago too
(as in your next quote). I suspect this depends on the University and
its resources. While our network is pretty slow at times its pretty
fast most of the working day. Once Universities decide to use
telematics-based resources routinely they'll either install a
purpose-built network or upgrade and police the one they have.

>12. Broadband is often supposed to allow the ugly duckling Internet to grow
>up and become the television medium it's always in its heart wanted to be.
>However, when @Home first rolled out its broadband service in California
>three years ago, it found that many of its subscribers were not interested
>in broadband interactive entertainment, but were rather @Work
>setting up >web servers on their home computers.

This is all about 'sit back' and sit forward' technologies. Its also
about using new media in the same way as the old and failing
miserably, a familiar tale which says nothing about how effective the
model could be.

>13. Broadband is definitely a wave of the future, but it has been hugely
>over-hyped and faces significant obstacles before it becomes a reality for
>the average user.

Judging from discussions at MILIA and VFX the over-hyping comes from
non- technologists who really don't understand the issues. The
internet, particularly the Web, is a new medium. It has two features,
1, the ability to support multiple media modes in all sorts of ways,
2, a mis-match between local computing power and bandwidth in the
form of lots of local oomph but inadequate bandwidth (which may get
worse before it gets better). Your quotes suggest that this will be
with us for a long time to come (but my comment about reorganising
what we have suggests maybe not so long). The key really lies in
being able to match demand with supply, which isn't happening yet.

The point about telematics-based learning isn't that you can put a
server at X and the whole world comes running to you, because as your
quotes make clear not only is the bandwidth not there but your server
has to be able to meet peak demand which is unpredictable.
Distributing the resources which manage your course makes the best
sense, however its done. More importantly you can deliver your
lessons in the style the student prefers at the rate he prefers and
at the time he prefers. What's more you can do it in a style which
would be all-but-impossible for a low-cost but endlessly repeated
lecture to replicate, mainly because the economics are different.
This is all in the paper I sent you before Christmas. (Is it up on
the web BTW - if so which paper is it? I would prefer to refer people
to it at the CIRCUS website rather than sending my own copies.) Some
really sophisticated styles don't require huge bandwidth but soak up
local processing power instead.

We don't know how to deploy this technology. The pedagogy simply
isn't there. I was hoping you'd have found some evidence that
students learn less well off telematics-based media, but all the
evidence I've seen is that they generally do better. This is the sort
of argument weakens my position that we really need high production
value presentations (yes, which include streaming video) but my gut
instinct is that a mixture of classical lecturing backed up by
automated reinforcement will be far more effective with high
production values than with the sort of text-based, low imagination
stuff we have on the web now.
The web can support it if the bandwidth used is local, but it is an
interesting exercise in the medium shaping the message. Again my
belief is that the limitations can be worked around fairly easily -
if you establish new models of usage.

John P.

   John Patterson    Tel:  +44 141-330-5323
   Dept of Computing Science  FAX: +44 141-330-4913
   University of Glasgow   email: jwp@dcs.gla.ac.uk

#  distributed via <nettime>: no commercial use without permission
#  <nettime> is a moderated mailing list for net criticism,
#  collaborative text filtering and cultural politics of the nets
#  more info: majordomo@bbs.thing.net and "info nettime-l" in the msg body
#  archive: http://www.nettime.org contact: nettime@bbs.thing.net