role of the network in the organization is increasing with the added dimension
of the internet. With intranets and the internet, the company network today is a
possible substitute for the sales and marketing functions. Typically, the
spending on management information systems by US companies has been 1% of their
revenue. An equivalent figure for sales and marketing has been 25-35%. If the
networks take over this function, the spending on IT will increase tremendously
over the next few years. Another way of looking at this is, if something like
50% of the total selling price of any product goes to manufacturing, the balance
goes to distribution. If the net takes over distribution it again means
considerably higher spending on IT. And in the US alone, the retail business
itself is worth over $1.5 trillion annually.
As this happens the end user will
want to use only the simple browser on his desk. The real and more complex
technology will go into the network. Just as the electronic circuitry moved into
the chip, the computer will now move into the network. And as networks become
immensely powerful they will reward excellence. Good news will travel incredibly
fast rewarding the winners. But so will bad news. This means mediocrity will get
harshly punished. And because the internet is a global phenomenon, it will
unsettle economies which have so far worked on the basis of isolation. In many
ways the fundamentals of economies are already being challenged. Earlier when
people earned more they spent more, and this led to inflation since the methods
of buying and selling did not change. Now the methods of selling are changing.
You can buy more and more for lesser and lesser. Therefore, inflation gets
contained even as earnings go up.
To make this
economy happen you need two things–computing and bandwidth. Computers have
progressed a lot in the last few years, but networks and bandwidth are in an era
which the mainframes were in a few years back. Mainframes required large
investment, long periods of return on investment and greater lock-in periods for
investments to pay off. The PC has changed all that.
However, really elaborate
networks work the same way even today. A 2.5Gb pipe between Los Angeles and Las
Vegas may still take six months to set up, cost $200,000 a month and need a
10-year lock-in period. This worked fine till voice became the key usage. Now
one multimedia transaction may need the equivalent of 10,000 phone calls. Again,
if the networks have to deliver transactions in the commercial workplace you
need bandwidth that can help you buy what you want where you want it. The
challenge for service providers is therefore not only to build bandwidth but
also be able to effectively distribute it when and where it is desired. This
also needs networks which work in an evolving manner, are cost effective and
flexible, and are scalable as per customer needs.
The network now has to work in a
transactional dial tone mode, like the telephone instrument which works when you
lift the handset–you get the dial tone and you pay for usage. Bandwidth has to
work very much the same way. Available when you want it, unlimited for the
duration that you need it and paid for on a transaction basis with no lock-in
periods for years for the purchasers of bandwidth. It has to become a tradable
commodity where the creators of bandwidth sell it through intermediaries to
those who want it at any point of time.
Make networks intelligent
intelligent optical networks. Networks that will be able to process much higher
amounts of data traffic in an intelligent manner with efficient network
management. Today’s networks do that but only for the voice domain–not for
the increased data traffic of the future. This will need both hard optics and
soft optics. Hard optics in the form of optics-based equipment that push in more
traffic through pipes without constant repeat cycles and noise distortions. Soft
optics in the form of software that will provide traffic management required at
the access, transport and switching levels of network management. Electrical
signals will have to be converted to optical ones and transported through fiber
channels over long distances taking alternative routes when required. The ideal
situation, of course, would be not to need to convert at all, but that needs
optical computing. And that is still 5-10 years away.
The interim answer
In the meantime,
existing networks have to be beefed up in a non-intrusive way. This can be done
by providing intelligent optical equipment that cut down the need for
regeneration of signals along the way and make them travel longer distances.
This also provides for routing of optical signals along paths where bandwidth is
available in real time. And most of all it helps increase the existing bandwidth
on optical channels at a much quicker pace. Thus, carriers of information will
be able to deliver bandwidth on a per transaction basis rather than on a fixed
The new revenue opportunities in
the optical framework today are the point-to-point optical wave services,
optical virtual private networks and gigabyte-level wave services. These will
need new blending of hardware and software solutions, which over the next 12-24
months should be able to deliver cost-effective solutions for the explosive
growth that data networks are going to need if the internet economy has to
In summary, intelligent optical
networking will reshape economies of the global public network, transform raw
capacity in the network into usable, economical and manageable bandwidth for the
delivery of new services, and enable a new generation of high-speed data
services over networked lightpaths.
report from New Orleans, Louisiana, USA, and New Delhi.