Sazia Mahfuz
Department of CSE, East West University, 43 Mohakhali C/A, Dhaka-1212, Bangladesh
Keywords: Grid approach to voter registration, Distributed election process.
Abstract: Bangladesh is a developing country which has gained independence in 1971. In the past, voter registration
was done by going from house to house and then filling up the required forms and counting them in. So any
type of computerized information system was not used. Currently, the government has taken initiatives to
collect voter information with photograph to effectively execute voter registration. This is definitely a
positive initiative from the part of the government. In this paper, to effectively limit the time duration and
complexity of the ongoing process, grid computing infrastructure implementation is proposed. Grid
infrastructure will not only accelerate the voter registration process, but also speed up the election process
on the whole.
The bangladesh election commission (bec) is a
constitutional body responsible primarily for the task
of conducting elections for the parliament and the
local bodies ( A significant
part of the task is the preparation, maintenance and
periodic updating of an electoral roll conforming to
the norms of the structure.
Recently the BEC faced significant difficulties in
discharging its responsibilities due largely to the
process adopted in preparing the electoral roll. The
9th Parliamentary Election was supposed to be held
on 22 January 2007 following the dissolution of
Parliament in October 2006. But, in the days leading
to the elections, violence became widespread with
deterioration in overall law and order situation of the
country. Eventually, the elections could not be held.
A new caretaker government assumed charge and
vowed to the nation its intention of holding a free,
fair and credible election.
One of the main reasons for the deferment of the
January 2007 election was a faulty and grossly
inaccurate electoral roll that was not acceptable to
the opposition parties. To improve the situation, the
newly reformed BEC took steps for preparing a
credible electoral roll with photograph. The attempt
has been appreciated to all parts of the people of the
country. In this paper, an approach to speed up the
current process further has been proposed.
After the independence, Parliamentary elections
were held for the first time under the new
constitution on 7 March 1973. Later in the year,
elections to local bodies were also held
(“Banglapedia”,n.d.). According to the article no.
118 of the constituency, there shall be an Election
Commission for Bangladesh supervising the whole
process of elections. Up until now, eight
parliamentary elections were held. In all of the
election processes, the voter registration process was
somewhat manual in the sense that the listing was
done by going from door to door and keeping track
of the voters in a written list. There the enumerators
designated by the Election Commission take account
of all the voters under his area. He hands down the
list to his supervisor, who after further scrutiny
submits it to the Assistant Registration Officer
(ARO). Then the list is counted and the proper list is
maintained in every polling centre.
This process had several drawbacks:
A voter can cast multiple votes in different
polling centres under different names as there
weren’t any biometric identification.
Fraudulent voter entry can be created due to the
lack of proper identification.
Mahfuz S. (2008).
In Proceedings of the International Conference on e-Business, pages 480-483
DOI: 10.5220/0001912004800483
As a result, outcome of the election process had
become unreliable and fraudulent. General people
had lost trust in the election process due to the
inefficient techniques adopted for the voter
registration process. This led to the unrest in
January, 2007.
An improved system has been achieved by
developing a system whereby the computerized
database is generated and is proposed to be
maintained through annual update using a process of
addition, deletion and transfer. This exercise
involves the capture of photographs and the
integration of these photographs with data in the
computers. The whole process can be summarized
through the following phases
3.1 First Phase: Data Collection”
Enumerators are provided with the appropriate
forms and required items using which they collect
data from the prospective voters. They have to keep
records of the mentally or physically challenged
people as well. The information is then handed over
to the supervisors and after scrutinizing the data, it is
handed over to the Assistant Registration Officer
(ARO). The data is compiled according to the areas
after scrutiny by the ARO.
The voter registration centres and area wise
server stations are then set up with assistance from
army. The registration forms are provided to the
team leaders by the AROs. The team leaders then
inform AROs of the data, time and numbers for
photograph taking. The information is then again
passed on to the enumerators from the AROs
through the supervisors. The enumerators inform the
prospective voters about the time and date on which
they have to be available in the voter registration
centre for photograph taking. The necessary
preparation is taken in the voter registration centre,
i.e. the laptops, webcams, fingerprint scanners,
photocopy machines and generators are set up for
the operations to be performed for data collection.
3.2 “Second Phase: Re-scrutinizing the
The forms maintained at the registration centres are
scrutinized again and then entries are created in
logbooks after serializing the forms in laptops.
3.3 “Third Phase: Data Collection and
Identity Scrutiny at Registration
Voters provide their information to the data entry
operators after scrutinizers have verified them. The
entries are created if required and completed after
taking the fingerprints, photos and signatures from
the voters. Then the voters are given receipt.
The information thus procured is again handed
over to the team leaders using laptops and they are
transferred to the area wise server stations.
3.4 “Fourth Phase: Registration of
Physically Challenged. Jail
Inmates, Missed-out Voters and
Ailing People”
After collecting data from the regular voters, all the
other prospective voters who have missed out the
registration have to be taken into account. They have
to be counted for following the steps described in the
earlier phases. The registration for the physically ill
people and the jail inmates have to be completed as
well, under the required arrangements.
3.5 “Fifth Phase: Data Processing at
Area Servers”
As all data have been collected for the voters, the
information has to be processed in the respective
area servers. The duplicate voters have to be
identified and the team leaders are informed. Then
the errors are corrected as required. The national ID
cards are processed for distribution after identifying
and correcting the errors in the exhibited draft
national ID cards.
3.6 “Sixth Phase: Draft Voters List
Printing and Finalizing”
Finally, the draft voters’ list is prepared and
exhibited. After hearing the objections on the list,
corrections are made by revising the proper
authority. Then the final voters’ list is prepared and
authenticated by the registration officer.
The electoral roll with photograph thus prepared
is proposed to be merged into a central database by
incorporating the fields necessary for voter
registration. This initiative has been laudable
compared to the earlier system that had always been
used for voter registration. But will it be successful
or not, that is yet to be seen. A better approach to
implementing this scenario will be grid computing
Figure 1: Physical movement of information in the fifth
phase of the existing scenario.
The ongoing voter registration can be further
improved with respect to the required time. The
proposed plan is to implement grid infrastructure for
the voter registration process. Currently, the voters’
information has to be transferred to the area servers
from the voter registration centres for data
processing as described in the fifth phase. Our
proposed plan suggests a change of the architecture
used in this phase.
The voter registration centres already have
laptops, fingerprint scanners, webcams for taking
information from the voters. Every voter registration
center is also connected to the existing public
switched telephone network (PSTN) provided by
BTTB (Bangladesh Telegraph and Telephone
Board) ( So if there is an
online interface for connection to the grid computing
system (Foster, 2006), then the need for the transfer
of information from the registration centres to the
area wise servers will be eliminated. All the voter
information taken in all the voter registration centres
can be directly fed into the proposed grid
This approach will obviously limit the time
duration required for the data transfer into the area
wise server and data processing required in those
server centres.
Moreover, it will pave ways for many more
opportunities for e-government related services.
Figure 2: Proposed scenario for Voter Registration in grid
implemented approach.
According to the roadmap (
provided by the Bangladesh Election commission,
the process of enumeration, data entry, printing,
posting and correction of the draft list will start from
7th August 2007 and finish on 16th August 2008. So
this phase of the process will have an approximate
time span of one year. This extended time duration
is due to the lengthy process of data collection and
then transferring the information to the area wise
server for re-evaluation.
This time span will be substantially reduced if
grid infrastructure is implemented. A simple online
interface can be developed to use the distributed
access points. The voter registration centres
distributed all over the country can be used as the
access points for the grid. The components required
to maintain the grid can be distributed in any number
of servers located at the different government bodies
secured from outside intervention. The need for
manpower in local area wise server maintenance can
be removed through this approach.
“A gap exists for all e-government projects between
the design assumptions/requirements and the reality
of the client public agency. The larger this gap
between design and reality, the greater the risk that
the project will fail (“DRG”,n.d.)”. The technique
Registration Center
at Barisal
Registration Center
at Narayanganj
Registration Center
at Khulna
Voter Registration
Center at Barisal
Voter Registration
Center at
Voter Registration
Center at Khulna
Area Server
at Khulna
Area Server
at Barisal
Physical Movement of
Area Server
ICE-B 2008 - International Conference on e-Business
followed in design-reality gaps model (“DRG”,n.d.)
presents a set of questions which will attempt to rate
the gaps.
The DRG model involves questions relating to a
series of seven 'ITPOSMO' dimensions -
information, technology, processes, objectives &
values, staffing & skills, management systems and
structures, and other resources - with attached rating
numbers (“DRG”,n.d.). Using each of the seven
dimensions, two things have to be observed. One is
the organizational reality that exists at the time of
analysis and the other one is the requirements within
the design of the proposed e-government project.
For each one of the dimensions, a numerical rating
has to be allocated to indicate the size of the design-
reality gap on that dimension. The rating for each
dimension's gap can be anywhere on a scale from
zero to ten.
After the rating, the rating numbers for all seven
ITPOSMO dimensions are added up and can be
interpreted according to the table 1 (“DRG”,n.d.).
The seven dimensions are applied to the
proposed approach and the corresponding rating is
Information: The information currently being
used for voter registration is the same as required by
the approach. So no change indicates 0 rating.
Technology: The required software for the
development of the grid computing environment
would be slightly different than the setup currently
being used. So a rating of 4 can be allocated to this
Processes: The work processes undertaken in
the current scenario need not be increased for the
requirement of the proposed scenario. So a rating of
0 can be assigned to this dimension.
Objectives & values: The objectives and values
required for the key stakeholders of the proposed
system have to be exactly the same as the objectives
and values of the current system. So a rating of 0
should be assigned corresponding to this dimension.
Staffing & skills: The staffing members and the
skill level or type will slightly differ in maintaining
the control of the grid architecture. So a rating of 3
has to be assigned for these criteria.
Management systems & structures:
Management systems and structures required for the
implementation of the proposed approach will be
slightly different than the exiting system. But the
overall management will be the same. So a rating of
4 can be assigned for this dimension.
Other resources: The time and money required
to implement the proposed system will be higher for
the initial setup. But it will definitely be lower for
maintenance compared to the existing system. So a
rating of 2 can be assigned.
The total rating for the ITPOSMO dimensions is
added up to 13. According to the table I, there is a
high possibility of success in the proposed approach.
Table 1: Overall rating analysis.
Likely Outcome
57 – 70
The e-government project will almost
certainly fail unless action is taken to
close design-reality gaps.
43 – 56
The e-government project may well fail
unless action is taken to close design-
reality gaps.
29 – 42
The e-government might fail totally, or
might well be a partial failure unless
action is taken to close design-reality
15 – 28
The e-government project might be a
partial failure unless action is taken to
close design-reality gaps.
0 – 14
The e-government project may well
Grid computing has become the popular solution to
distributed problems. Election process is a very
critical distributed problem faced in any
government. In this paper, a better approach is
proposed to maintain the database of the voter
registration, so that the distributed election process
can effectively be handled with the least possible use
of centralization.
Foster, I.(2006). Globus Toolkit Version 4: Software for
Service-Oriented Systems, IFIP International
Conference on Network and Parallel Computing,
Springer-Verlag LNCS 3779, pp 2-13. (15th April 2008).
Retrieved: April 15, 2008.
Banglapedia:history (n.d.). Retrieved: 17th April, 2008, Retrieved: 17th April, 2008.
DRG Model
shtml . Retrieved: 27th May, 2008.