by November 28, 2000 0 comments

This article discusses a set of products and services that
will be used in large organizations. Before we get into the trends bit, let us
first define the Enterprise. There are as many definitions of the enterprise as
there are people attempting the definition. For the purposes of this discussion,
we will define the enterprise as any organization that has a physical presence
in more than one geographical region and which requires its IT infrastructure to
be available twenty-four hours a day, seven days a week.

Heterogeneous network

Increasingly, enterprises, and even smaller corporates, are
using heterogeneous networks. That is, they are mixing and matching operating
systems (OSs) on their network servers, instead of having one OS on every

Initially, the case for deciding on any one OS across the
network was very strong, particularly from the point of view of interoperability
and system administration. But today, almost no one will subscribe to that view.
There are two major reasons for this. First, migrating from one OS to another is
an even more difficult and time consuming task than making them interoperate.
Second, it is an accepted fact today that each OS and the applications that run
on it have their own areas of strength. And no single OS tops in all network
requirements. Combined, these two factors are ensuring that enterprise networks
are strongly heterogeneous.


Storage Area Networks (SAN) are the direct result of the
explosion of the number of servers on the corporate network, each with their own
terabytes of data. Managing all this can be quite a nightmare. With SANs, the
storage part is separated from the servers and set up on a separate system, with
high speed (fiber channel) connects to the rest of the network. This way, the
server is left to handle the applications it is running, without having to
bother about managing the data also.

Not every Tom, Dick, and Harry will need, or be able to
afford, SANs. SANs are for those who have a huge amount of live data on their
servers, running into terabytes. Also, the bill for setting up a SAN can be
pretty stiff, running into a couple of crores just to get started. So you know
why there are not too many SANs around.


In the world of the Application Service Provider (ASP), you
would not buy applications–you’d rent them over the Net. Again, you will not
rent full applications, but only those parts that you need to use. For example,
if your need is word processing, you will not buy Microsoft Word or Corel
WordPerfect. Instead, you will rent a word processing application. And while
renting it, you will not rent the thesaurus if you do not want to use it.
Basically, the service provider will set up and maintain the application. You
will access it over the Web, ideally on your browser. Your data could also
reside in the servers of the ASP.

ASPs were supposed to become big business this year, but did
not. This is primarily because of paucity of bandwidth. You need quite a bit of
bandwidth to run an application off a server and store and access all your data
there. We are not yet ready with all that bandwidth. Even where the bandwidth is
there, concerns of data security kick in. Not everyone is sure that they can
entrust their valuable data to someone else to store. Also, there has to be
significant savings in cost or effort for corporates to switch over from the
current owned model to the ASP model. ASPs are still in the process of
convincing corporates that that indeed is the case.

Call center

Customer support systems are becoming critical elements of
business success. For multi-outlet businesses like banks, chain stores,
e-businesses, etc, a centralized support center is fast becoming a must. A call
center is a place where customer telephone calls are routed to, and answered.
The advantages of having a centralized system instead of taking the call at
individual shops are many. To start with, you can install sophisticated customer
support systems and information databases far more easily at these locations, as
opposed to having them at multiple smaller locations. Call centers are rarely
set up at the business premises itself. Telephone calls can easily be switched
across the globe. So, you can choose a location that is the cheapest.

Today, many third-party call centers are coming up,
supporting multiple businesses off the same location. So, you need not go
through the trouble of setting up your own. Like with medical transcription and
offshore software development, India is emerging as a potential site for
offshore call centers.

Data center

A data center is something like an ASP, without application
rental being the primary focus. You use the services of a data center to host
your data off site. Data centers in their simplest form are huge server farms
with very fat pipes connecting them to the Internet. The most common use of a
data center is for Web hosting. So far, Indian companies were hosting their
Websites at data centers abroad. This was not because servers were not available
in the country, but because the required bandwidth was not available. As
bandwidth availability improved, we graduated to co-locating our Web servers.
That is, we put the Web server at the ISP’s backbone itself. A further
refinement kept the server at our own premises, with a fat leased-line
connection running to the ISP’s backbone.

Data centers are the next stage in this evolution. And they
have just started making their appearance. Today’s data centers are not just
Web hosting centers. In fact, they offer a multitude of services to their
clientele. For starters, many data centers do offer or plan to offer ASP
services. Then, they can manage all your data on their servers, instead of your
having to bother about backup schedules and anti-virus updates among other
things. In short, data centers can well be the next stage in the evolution of
the annual maintenance contract. What started off as hardware maintenance
contracts now include on site management of your network, and can tomorrow
evolve into the complete off siteing of your data. What advantages will this
have? The argument is that you do what you are good at–your business–and
leave those who are experts at managing data to them, as you are doing with your
systems today.


Setting up a Virtual Private Network (VPN) is a way of
connecting networks at different geographic locations into one, over the
Internet. Consider a company that has offices in multiple locations. Before VPNs
happened, the only way of connecting up these networks was to go in for leased
lines. VSAT connections were few and far in between. Leased lines are a costly
option and the farther the locations are, the costlier they become. VPNs
provided a way out. Here, all your offices connect to the Internet and the data
is sent over the Net.

This way, the length of the leased line is limited to, from
the office to the nearest Internet access point. You do not need crisscrossing
connects going all the way across. This could have been done even a few years
back. But the question of data security came in the way of that happening. Then
came the VPN. The development of tunneling protocols ensured that your data
could be secure even while traveling over public networks. And finally there was
a way for smaller (and even larger) organizations to connect up their offices
into one large network without paying for it through their nose. Expect more
action on the VPN front in the years ahead.


Broadband is a hot topic, both for the home and the office.
Strictly speaking, broadband is discussed in the context of the home. Offices
are supposed to use leased lines. But that is not exactly the case in India.
Broadband options–DSL, cable modem, etc are being considered for offices too.
Why is it that broadband is not the primary option for the office? Broadband
technologies are mostly one-way technologies. That is, downstream speeds are
many times the upstream speeds with these options. In a home or individual
setup, more data is expected to flow from the Web down to the PC you are using,
while you will send back only a limited amount of data–e-mail, requests for
pages, etc. Unlike this, an office needs to send large amounts of data both
ways. This is partly because offices have many users connected and they need to
send a lot of data both ways over their Net connects. For example, if you have a
VPN running, or if employees need to frequently update portions of your Website,
etc, the ideal option for you will be a connect that offers the same data
transfer speeds both ways–something like a leased line.

For purely Web browsing needs, enterprises can opt for a
broadband solution, but pretty soon you will be looking for other options.


This is actually old news. Most large enterprises have
already been there and done that. Smaller organizations have also done that.
What makes ERP interesting again is that vendors are trying to move ERP to the
Net as an ASP offering. That is, you do not need to go through the rigors of
implementing an ERP solution and then maintaining it. All that gets taken care
of by the ASP. Theoretically, you can be up and going right from day one. This
model is supposed to be advantageous for both the small organization and the
large enterprise.

ERP applications are a hot prospect for ASPs. And all the
factors that are holding back generic ASPs are holding back this one too.

Data mining

Like ERP, data mining is also old stuff. But not too many
corporates in this country are known to be doing this seriously. A significant
part of the reason is the non-availability of details, like detailed buyer
profiles, etc. What is data mining? It’s simply taking available data and
searching in it for patterns that you can use for business advantage. For
example, consider the simple case of bills being raised and payment being
collected. Assume that you have collected a year’s data. If you analyze this
data carefully, you can make out patterns like, in which areas are your bills
paid faster, or clients in which age group and educational background are more
likely to default on payment.

Does the example sound familiar? Like something that you may
actually be doing? Data mining can be considered to be a highly evolved form of
good old MIS reports in action, where you work with much larger volumes of data
and more intelligent tools. Also, you are looking for patterns as opposed to
numbers. Obviously, the more the data (quantity and spread) you have, the better
the patterns will be.

Krishna Kumar

No Comments so far

Jump into a conversation

No Comments Yet!

You can be the one to start a conversation.