December 22 / December 29, 1997
Web Bandwidth Problems Need More than Band-Aids
By Bob O'Donnell
Another year is upon us, so it's time to start making predictions for 1998. I'm going
to go out on a real limb and predict that Internet bandwidth limitations will be a big
issue next year. I know, I know -- this isn't much of a stretch. But I think lots of
people are going to be begging, nay, demanding faster connections to the Internet over the
next 12 months.
There are several important developments scheduled for 1998 that should help both
individual users and large sites. First, the official adoption of the V.PCM 56K modem
standard in February will undoubtedly lead to rapid, widespread adoption of 56K modems and
the creation of 56K dial-in lines from virtually all ISPs. This one development, in and of
itself, will be great news for mobile users, telecommuters, and even lots of small
offices.
Second, the formal approval of the HTTP 1.1 protocol (which is currently supported in
both Netscape Navigator 4.0 and Internet Explorer 4.0), should lead to a more wide-scale
deployment across Web servers and other pieces of the chain that links end-users to
information. HTTP 1.1 offers many improvements, the most important of which seems to be
more efficient use of the connections between browsers and servers. To end-users, this
ultimately translates to speedier downloading of Web pages, regardless of the size of the
pipe you're using to connect.
And, organizations that are well-enough funded and organized will next year implement
fast networks, such as Gigabit Ethernet 100 Mbit Ethernet, which may help corporate
networks deliver faster connections to individuals.
Many of the other proposed "solutions" to Internet bandwidth seem more like
Band-Aids to the problem, however. Products such as PeakSoft's PeakJet and
forthcoming Net Magnet,
and Connectix' new SurfExpress,
for example, attempt to speed up Web browsing by creating large personal caches,
pre-fetching pages off the Web into these caches, and other similar types of tricks.
I can certainly understand the theory here. L2 caches have become an indispensable part
of today's computer systems. Still, although I don't have any first-hand experience with
these three, I have to wonder how successful the notion of a personal proxy server can
really become. Certainly in a large corporate environment, these bandwidth-hogging
packages are bound to be nixed as quickly as PointCast and other "push"
products.
The real issue -- the size of the pipe -- is what needs to be addressed. These products
just serve as a masquerading distraction.
Given how desperate many people are to try anything that offers faster access to the
Net, I'm sure we'll be hearing a lot about and seeing many different types of these
personal proxies, or other similar tools. In fact, I predict that this will be one of the
more active new segments in the software industry in the coming year. As far as I can see,
though, the outrageous claims about these products' capabilities will prove to be much ado
about nothing.
On a cheerier note, I hope that you and your families have a joyous holiday season.
Merry Christmas, Happy Hanukkah, Happy Kwanza, and Happy New Year. I'll see you again in
'98.
©
Copyright 1997, by InfoWorld Publishing Corp., a
subsidiary of IDG Communications, Inc. Reprinted from InfoWorld,
155 Bovet Road, San Mateo, CA 94402. Further reproduction is prohibited.
