Information Technology Reference
In-Depth Information
4.5.1 Main Enterprise Network
The main enterprise network contains all information that is easily accessible. In our
pilot project most of this includes the city's web pages where information about the
city itself is presented to the general public. In addition, there are also intranet pages
beingmaintained to help employees in their dailywork. For example, a list of contacts
and their work division is published on the website. These lists are very important
for the employees, e.g., when they want to arrange a meeting with a colleague for a
particular purpose. Summarizing, web pages both in internet and intranet represent
a common platform to represent the city of Berlin either externally or internally.
Consequently, these are valuable resources that should always be accessible for the
employees. Accessing this information usually requires no additional authentica-
tion since both extranet as well as intranet are usually accessible for anyone whose
computer is located in the main enterprise network.
In order to provide access to the web and intranet pages for anyone within the main
enterprise network, we deploy multiple crawling ensembles, consisting of crawling
and search agents. Therefore, we utilize open source tools Nutch
4
and Solr
5
and
integrate them with our implemented software agents. This enables us to customize
and automate the crawling and indexing process for vast amounts of data. In the Berlin
administration office, employees' general information (i.e., their names, departments,
email addresses, and telephone numbers), i.e., the administration's meta directory,
is maintained using an LDAP server. For crawling such lightweight data we rely on
Lucene.
6
All of these crawling processes run independently, creating different disjoint
indices in the process. A search agent is responsible for one of these indices. Search
requests are forwarded from the broker agent to these search agents. The broker agent
is also responsible to merge the search results into a single result list. This result list
is then presented to the user using the graphical user interface.
4.5.2 Department Network
The bigger the size of an enterprise, the more complex its hierarchical structure will
be. It is not unusual that companies are split into different divisions or branches.
From an IT infrastructure point of view, these subdivisions often have their own
subdomains within the main domain of the company or administrative division.
Since these subdivisions may, to some extent, work independently of each other, it
is not unusual that each has its own data administration and access policy. Besides,
each subdivision usually maintains its own file servers that should exclusively be
accessed by the employees of this particular division only.
6
http://lucene.apache.org/
.
Search WWH ::
Custom Search