Few phenomena in
history have matched the explosive and rapid growth of the world wide web.
Starting as a means of exchanging information among researchers, it has rapidly
developed into one of the largest communication mediums in the world.
Entrepreneurs were quick to see the potential of this new medium–the only one
among the mass media to possess the characteristic of two-way communication.
Like the other mass media, the web could make information available to a vast
audience. However, the unique ability to provide two-way communication implied
that the web could also carry back responses from the audience, thus providing
the ability to carry out commercial transactions. The ubiquity and speed of the
underlying communication medium promised to provide a single universal
marketplace that spanned the entire world.
Challenges
Quantitatively,
the growth of the web has been extremely rapid by any standards. The number of
users has increased from virtually nothing in 1995 to 200 million in 1999 and is
expected to double to 400 million by the year 2002. These 400 million users are
expected to generate about $600 billion in commercial transactions, an almost
tenfold increase in value from 1999. Achieving this impressive growth will
require overcoming a number of data management challenges.
ENTERPRISE DATA MANAGEMENT:
Fully automated ecommerce requires that enterprise data be consistent and
up-to-date. Unfortunately, this is almost never the case. Historically, each
department in a company maintained its own information, which was manually kept
consistent with the information in other departments. Computerization led to the
situation where each department became an ‘island of automation’, but
enterprise-wide functions could not be automated due to inconsistent data. For
example, the employee address may be maintained in both the human relations and
finance databases. In the case of a change, the employee may have the
responsibility of notifying both departments, or there may be an ad hoc
procedure for one department to notify the other. Untangling this maze of
overlapping information and keeping the information in these different databases
in sync is not a task for the faint at heart.
PERSONAL DATA IDENTITY
MANAGEMENT: The interactive
characteristic of the web poses a great danger to privacy. In other mass media,
such as TV, it is difficult to collect detailed personal data about individuals
by observing the material they watch, since such data cannot be collected. The
two-way communication nature of the web, on the other hand, makes obtaining such
data easy. Since most democratic societies agree that some degree of privacy is
desirable, this is a difficult problem. For instance, most individuals would
agree that it is undesirable for strangers to guess that one may have a health
problem by observing the fact that one is visiting websites dealing with health
information.
The challenge of identity
management has two aspects. The first aspect is that each website can, by
analysis of usage patterns, obtain information about individuals. The second
aspect is that websites can pool information to obtain greater information about
individuals than these individuals could realize.
SYSTEM AND NETWORK MANAGEMENT:
As the internet becomes pervasive, the number of workstations and servers in
enterprises will increase. As a result, the cost of maintaining these machines
will rise. Some studies have indicated that the annual maintenance cost of a
computer is approximately equal to the combined price of the hardware and the
software. That is, the cost of maintaining a system costing around Rs1 lakh is
Rs1 lakh per year! While this may be extreme, there is no doubt that the
maintenance cost of a system is significant and it would be ironic if the
productivity gains from using the internet were to be lost due to the increased
cost of maintenance.
Directories: Yellow Pages of
the network
The closest
analogy to a directory service in everyday life is the Yellow Pages in a
telephone directory. The Yellow Pages may contain a list of businesses, divided
into various classifications such as florists and restaurants. A classification
can be divided into sub-classifications. For example, restaurants can be
sub-divided into Indian restaurants and Chinese restaurants. These could be
further sub-divided into south Indian and north Indian restaurants. The listing
for each restaurant would contain information about the restaurant.
In a similar manner, an
enterprise directory may contain information about all the entities–users,
computer systems and printers–present on the network. At the highest level,
the information is typically sub-divided by geographical location. At the next
level, the information can be further sub-divided by organizational units, such
as department or group. At a lower level, there would be entities in the
enterprise, such as users, together with information about the entities, such as
the address of the user or the computer systems they can access. Different types
of information can be stored for different types of entities–for instance, it
is not necessary to define a postal address field for computer systems.
Additionally, directory services provide a feature not provided by Yellow Pages–users
are restricted to only the information they are authorized to access. In other
words, unlike the Yellow Pages, a directory service stores information securely
and offers only secure access.
Though such features seem simple,
they are powerful. The key features of a directory service are:
-
the ability
to store structured information about entities. -
the ability
to make information available even over a widely dispersed geographic area. -
fault-tolerance.
-
interfaces
specialized for searching for information about entities. -
internal
organization specialized for searching for information. -
scalability,
both in terms of the amount of information that can be stored and the number
of searches that can be made every second.
These features can be summarized
by saying that an enterprise directory should have the capability to make
enterprise information available enterprise-wide.
How can directory services be
used to solve the problems described earlier? In all the solutions, the ability
of directories to make information securely available over the network, and
provide information in response to searches are important.
ENTERPRISE DATA MANAGEMENT:
A typical enterprise today suffers from information indigestion, due to the
multiple overlapping and inconsistent databases present. Some commercially
available directory services offer meta-directory services that can be used to
solve this problem. Meta-directory services allow a directory server to keep
specialized information about other directory servers in addition to information
about entities on the network. The directory service is notified that the
employee address is duplicated in the databases. It will then use its
consistency-enforcing mechanisms to ensure that any change made to any one of
the databases will be propagated to the others.
IDENTITY MANAGEMENT: A
fundamental concern today is that the world wide web can rapidly lead to a
complete loss of privacy. The problems are that websites can infer information
about individuals from the websites they visit, and that by pooling information,
websites can deduce more information than most individuals realize they are
giving away. Since the major function of directories is to store information
about entities, including individuals, they have a crucial role to play in the
management of identities. Privacy-enhancing technologies allow individuals to
control the amount of information they reveal, or even take on multiple
identities. A directory would need technology that allows users to define
multiple identities–referred to as profiles–in the directory. These
identities can be completely different from the individual’s true identity and
contain only the information the individual wants to reveal for that identity.
For example, an individual may define a personal identity for visiting websites
of personal interest and a professional identity for other sites. In the
personal identity, the user may choose not to reveal any income information.
Whenever the individual visits a website, an identity can be selected. Since
this identity can be completely different from the individual’s real identity,
it is impossible to infer anything about the individual. If the website asks for
any information, like during registration, the directory service will reveal
only the information appropriate for that identity. Therefore, even by pooling
information from multiple websites, only the information that the individual has
decided to reveal can be obtained.
System and network management
Directories
can also cure problems caused by rapidly escalating system management costs.
This arises from the ability of directories to provide network-wide information.
Therefore, directories can be used to store policies, which define how to
configure arbitrary computer systems or subsystems. A policy can be used to
define the preferred desktop configuration, including applications and
wallpaper, of an individual user. Since directory information is available
network-wide, this information can be retrieved to restore this configuration on
any desktop the user logs in on the network. The directory service can also
ensure that all members of a particular department have access to a particular
set of applications by associating an appropriate policy with the department,
which is likely to be a sub-division in the directory. Thus the directory is
used to solve the major maintenance problem of software installation and
configuration. Additionally, an inventory control component stores in the
directory hardware and software components of all systems in the
enterprise. As a result, the directory will contain an up-to-date inventory that
can be used by asset management and financial analysis programs. The application
of these concepts is widening.
Dr DINAKAR
SITARAM