Sie sind auf Seite 1von 10

LinkedGov Government Social Network

and Software-Defined Workflow


Khanh Dao, Fuyao Zhang, Yu Cheng, and Yuexing Yin
Department of Software Engineering, San Jose State University
khanh.dao@sjsu.edu
fuyao.zhange@sjsu.edu
yu.cheng@sjsu.edu
yuexing.yin@sjsu.edu
st
Abstract - In the 21 century, with the rapid growth of social platforms such as Facebook, Twitter, and LinkedIn,
social networks have become an important part of our daily life. Social networks have established the
communication structures and social behaviors based on various factors, such as genders, ages, friendship,
common interests and career backgrounds. Governments have also discovered the potential to improve the public
and private communication, information sharing, and outreach. A government social network system is usually
known under the term Government 2.0 or e-Government. Today, the existing e-Government system faces many
challenges, such as untenable connectivity and permissions, manual and error-prone defined workflows.
Workflows, tasks and permissions are usually difficult to be tracked or monitored. Moreover, the software systems
are inflexible to adapt to change, and insecure in the way information is shared, which usually become the potential
targets for hackers. Accordingly, there is a need to improve and optimize the governments social network systems.
In this paper, we propose LinkedGov, the software-defined architecture, which is capable of creating and
maintaining multi-layered social network topology. The approach enables a feasible hierarchy of connectivity,
permissions and services; enhances security and flexibility in managing workflows amongst different governments
social groups and departments.

Index Terms e-Government, social network service, workflow management software, task management software.

I. INTRODUCTION governmental data and information. The


access information varies from weather
In the past decade, with the growth information, to accidents and criminal
of social media platform and online data. This requires data should be more
community, social network has become accessible and up-to-date, which
new tools for public sector to requires a better communication and
communicate publicly and internally, in interaction between departments and
order to improve the transparency, members inside the government. This
participation, and collaboration. E- also proves the huge need for more
Government has aggressively promoted modern solution that is able to respond
the new provision of government, where to citizens requirement for quality
workflow and process could be public services, which are fast, reliable
integrated and executed securely and and accessible. An important feature any
seamlessly. There is also increasing e-services is the interactivity: e-
needs to for governments data with Government system should provide
better metrics in order to study the citizens tools to activate process of
modern society for future prediction and service delivery anytime and participate
development. According to the recent actively. There is a significant potential
report, as of 2015, 76% of Americans for technology in this market, with many
have used the Internet to access different sections need to be covered.

1
and distribute from the given
network. In random graph, edges are
placed at random; therefore, the
majority of nodes have
approximately the same degree in the
network.
2) Scale-free network (Meso-level):
Not all nodes in a network have the
same number or edges. Majority of
nodes have only a few links, but a
few nodes have a very large number
Figure 1: Statistics of users who access of links. This structure enables
government services scaling mechanism for the network,
which is absent from the classical
LinkedGov aims to create a flexible distributed network. Most networks
social platform for engaging and grow through the addition of new
managing the complicated hierarchy in nodes with link already presented in
government and its current process. The the system. In most real network,
platform offers an online web and link to a node with large number of
mobile solution to configure and manage connections -- known as
governments internal relationship and preferential attachment -- has
process, authorize explicit and implicit higher probability.
permission on significant process,
communicate on issues between public
with government, and notify on changes
between process in real-time. The
project bridges the gap in
communication and security in e-
Government; provides the all-in-one
solution platform that enables
government to communicate internally Figure 2: Random & Scale-free network
between different departments and
levels, and execute complex tasks or 3) Hierarchical network (Macro-
processes more accurately and securely. level): This is a part of scale-free
model family, which uses iterative
algorithms to construct a network
II. SOCIAL NETWORK model that combines the scale-free
STRUCTURES AND PATTERNS topology with a high degree of
The following network structures are clustering, in order to bring
analyzed: modularity inside a large and
complex network structure. For
1) Random distributed network (Meso- example, Figure3 shows the iterative
level): This framework provides the construction leading to a hierarchical
capability to generalize the network: from connected cluster of
restrictive independence assumption five nodes (a), the iterative obtains a
of micro-networks, small sub-graph network of N = 25 nodes (b) and N =
configurations model to decentralize 125 nodes network (c) through

2
replicas process. This process can be Small-world network could be used
continued indefinitely. [3]. as hybrid structure in transforming
social links. A study, which is
conducted in Complex Networks:
Small-World, Scale-Free and Beyond
(IEEE Circuits and Systems
Magazine, 2003), shows the example
of transforming highly cliquish
network, where nodes have many
degrees apart to a scattered random
network, where few nodes have same
connections, and pairs are few
degrees apart on average.
Figure 3: Iterative construction in Hierarchical
network [3] Clustering coefficient, which is a
measure of the degree to which nodes in
In theory, complex network structure is graph tend to cluster together, is an
usually classified into four types: [22] important concept in node clustering.
Research studies the characteristics of
1) Random Network: These graphs are clusters in both small and large scale,
constructed by randomly adding and classifies them into four different
links to a static set of nodes, and tend types: General, Star, Aggregation and
to have short paths between nodes. Spread graph. Star graphs social network
2) Power-law Network: In this services have smaller average cluster
network, the probability that a node coefficient than General graphs;
has degree of k is proportional to k-y, therefore, they have smaller degree of
y is called power-law coefficient. In cohesion. In an extreme star topology
fact, many real-world networks are network, one user can connect with all
constructed based on this concept, other users, and all other users are
such as Internal topologies, the Web, connected with only that one user.
social networks, and power grids. Aggregation topology has higher
3) Scale-free Network: This is a class average cluster coefficient, and a shorter
of power-law networks, with high- average path; hence it has higher degree
degree nodes connecting to other of cohesion. This type of topology is fit
high-degree nodes. to maintain an intense relationship in
4) Small-world Network: This type of social network. Spread topology is
network has small diameter and similar to General topology, but has
exhibit high clustering. In this longer average path length.
structure, most nodes are not
neighbors of one another, but most
of nodes can be reached from every
other node by a small number of
hops or steps. The algorithm assumes
that the distance between two
random nodes is L, and grows
proportionally to the logarithm of the
number of nodes N in the network.

3
communicate within the friend
network cyclopedically; only few
communications happen outside the
network.
3) Inclusive friend network type:
Members communicate within
friend network and many
General topology (C1) Star topology (C2) communicate outside network,
which result high coverage and low
aggregation ratio.
4) Independent friend network type:
This network has both low
aggregation and coverage ratio,
Spread topology (C4)
since members communicate
independently with the friend
Aggregation topology network in this pattern.
(C3)
Figure 4: Four types of social network topology

On average, the number of nodes and


links are used on each topology is shown
in the table below:

Figure 6: Social network communication


Figure 5: Average network indexes patterns [4]

Beside network topology, social network III. SOCIAL NETWORK


is also classified based on PRIVACY
communication patterns [6]. Research
studies different patterns to result the There always exists a tradeoff
ratio for aggregation (number of between functionality and user privacy,
since users need to reveal certain
comments for friends/number of all information to make use of the desired
comments) and coverage (number of functionality in social network. Security
friends who post comments/number of and privacy concerns in two areas: (1)
friends of blog user). user-related and (2) provider-related.
1) Partial friend network type: In this
A. User-related privacy
communication pattern, members
communicate with limited group of Privacy breached can be through a
friends, which result high deliberate act (snooping, hacking), or by
aggregation ratio and low coverage. accident (mismanagements of privacy
2) Parity friend network type: This settings). (a) Stranger can view private
network has high aggregation and information through misconfiguration
coverage ratios. Members for restriction or design flaws from

4
social platform service provider. (b) data (Ex: obscuring the link between
Managing and separating social network user and data sold through
context can be blurry in social network cryptography), while preserving
platform. As the consequence, other structures of interest in the
information is unable to be hidden from data. Another technique in
a specific group. (c) Disclosure anonymization is differential
boundary is one of the security concerns privacy, which makes user
in social network, since information is computationally indistinguishable
made more public than intended. Data, from most of the user in data set. The
which means to be disclosed privately to drawback of this technique is that it
another user, is made available to a is hard to definitely prevent recovery
larger audience. private information, since most of
the obscuring is done through
B. Provider-related privacy classical cryptography, and external
hints might be available for attackers
The privacy threat occurs through the
to exploit.
trust that user might put to the service
2) Decentralization: It can be applied
provider in social network -- a
to different degree, in order to
configuration system that has full access
prevent implicit trust being abused.
to any user-related data, including
In an extreme form of
browsing behavior and message log. (a)
decentralization, all traffic will take
Data retention issues, which violate of
place through peer-to-peer networks.
the temporal boundary: information,
In general, the more decentralized
which is supposed to be deleted, is still
the solution, the better the protection.
available and accessible in social
3) Privacy setting and management:
network systems. (b) Social platform
Social network platform should
employee has full access to data and
make information disclosure and
might take advantage of this. (c) Data is
privacy settings more gradual,
often anonymized, before being sold to
transparent and fine-grained, and
third party or generated for marketing
give user more controls over their
purposes, in order to protect user privacy
privacy settings. This solution is
and identification. However, it is still a
cheap to implement, but requires
threat, that user can be re-identified by
user awareness in order to reach the
looking for unusual points within
full potential.
friendship graph. Selling and marketing
4) Encryption: Through cryptography,
of data is also a conflict of the implicit
one can protect data securely.
trust in social network, since information
However, key management is a big
is used for different purposes, than
issue that should be paid attentions to
intended by the users.
in this method. Cryptography also
involves expensive operations, which
The following are five technology
might result scalability issues in the
patterns in privacy protecting:
large social network system.
5) Awareness, Law and Regulations:
1) Anonymization: Usually performed
The non-technical approach aims to
once before sales to third party, basic
enhance user awareness of the
anonymization involves removing
privacy issues, through established
any identifiable information from the

5
laws and social conducts. However, is applied when the assumption made
it lacks the power to enforce the for the simple merge pattern does not
change in the system automatically, hold. It deals with merge situations
and heavily depends on users own where multiple incoming branches
behavior while interacting with the run in parallel, by activating its
system. outgoing node after being activated
by the first incoming branch, then
IV. WORKFLOW DESIGN waiting for all remaining branches to
PATTERN
complete before it resets itself.
Workflow management is another 9) Structural pattern: Deal with
important feature in governments social arbitrary and implicit restriction of a
network. Workflow patterns are divided workflow.
into the following types: 10) Cancellation pattern: Used for a
workflow instance being cancelled.
1) Sequence pattern: Linear execution
of workflow steps, one action is The component-based workflow
activated unconditionally after architecture should encapsulate
another finishes executing. workflow features in separate
2) Parallel split pattern (unconditional components: system must be able to
routing): Execution is divided into adapt to change, and separate
multiple parallel threads that can run functional core from extended
independently from each other. functionality and customer-specific
3) Synchronization pattern parts. Micro-workflow system consists of
monitoring, human-computer interface,
(unconditional routing):
and three core components:
Synchronize multiple parallel threads
into one single thread of execution. 1) Process component: provides
In this pattern, workflow execution abstraction and implements an
continues once all synchronized activity-based model.
threads have finished executing. 2) Execution component: consists
4) Exclusive Choice pattern: Define of the implementation of the
multiple possible paths for workflow functionality to execute workflow.
to be executed. 3) Synchronize component: define
5) Simple merge pattern: Merge the dependencies within the workflow
possible paths, which are defined by domain.
preceding exclusive choice, assume
that only one path is taken and no
synchronization takes place.
6) Multi-Choice pattern: Define
multiple possible paths for workflow
to choose from. It is the
generalization of the parallel split
and exclusive choice pattern.
7) Synchronizing Merge pattern: Used
to synchronize multiple threads of
executions that are activated by a
preceding multi-choice. Figure 7: Workflow engine architecture
8) Discriminator pattern: This pattern

6
V. PROPOSED DESIGN How information is Random walks,
PRINCIPLE spreading? hits algorithm.
A. Social Network Interpretations
How to increase Bridge.
The study on different social network redundancy and
platforms and their common interconnectedness?
characteristics has been conducted with
the following questions:
How dynamic the Temporal
organization is? networks.
Questions Social network
approach
Figure 8: Social network interpretation

Who is the central Degree centrality B. Social network design principle


in the trust and
respect network? Social network has highly dense
connections inside, which means each
How do we rank Eigenvector node has many connections to many
our top performer centrality, page- other nodes. Harvard University has
individuals in the ranks. conducted a study based on different
organization? parameters of social network
characteristics, in order to build the
How do we put Closeness model, which display the essential
together a team in centrality features of social network and make it as
network? simple as possible [6]. The model
generates the rules to construct the
network: (1) first, the degree distribution
How to merge Homophily, should have a steep tail; (2) second, as
separated network? reciprocity,
mutuality, the size of the network is growing, the
transitivity. average of the path length also should be
growing but in a slow speed; and (3)
third, the model should also display the
What is the missing Structural holes high clustering, and show the
links between
supply and relationship between each vertices and
demand? dense internal connections. The
algorithm is divided into two processes:
(1) random attachment; and (2) implicit
How strong the Tie strength preferential attachment, which results
impact of the
advertisement from the following edges from the
effort? randomly chosen initial contacts. The
following steps are executed in sequence
to construct the network graph:
How segmented the Clustering
market is? coefficient, clique,
cohesive, 1) Create a root node for social
modularity. network.
2) Pick on the average random vertices
as initial contacts.

7
3) Pick on the average neighbor of cluster.
each initial contact as secondary It is important to understanding the
contact. characteristics of different type of
network topologies in order to detect
4) Connect the new vertex to the initial overlapping nodes and link communities
and secondary contacts. in a complex social network, as well as
5) Repeat steps 24 until the network construct and design nodes with unique
has grown to desired size. identities. Many algorithms with
statistical, information-theoretic,
When a new node is generated in the synchronization and dynamical
clustering approaches have been
network, it will connect its neighbor developed for the purposes [11]. Two
nodes, which have the shortest path types of community structures based on
length between each other. Following sub-network connection: strong sense
the principles of the proposed algorithm, community has larger internal
the duplicate relations between each connections than external connections;
weak sense community has internal
node can be avoided [Figure 9]. A highly connections equal or smaller than
dense connection network could also be external connections, but its internal
constructed based on this algorithm. connections are larger than the
connections between the sub-network.
The co-existing of strong sense and
weak sense community is a common
structure of a complex social network
cluster. For example, in [Figure.10],
nodes divided into three clusters
(different colors indicate different
communities): strong sense community
(cyan nodes). (b) Overlapping member
property: node d is a second type of
Figure 9: Growing process of social network
overlapping node shared by cyan
community and green community. [10]
C. Detecting cluster

The construction of the social network


model include the following elements:
1) Nodes: Represent members and
department in government social
network.
2) Links: Describe relationship between
Figure 10: Social network cluster
nodes, based on the relationship
type, services or functionalities are
dispatched and processed
accordingly.
3) Clusters: Describe the distinct
relationship group that every node
could be able to communicate to
each other internally.
Communication between clusters
should be through common link.
Figure 11: Social network cluster model
4) Graphs: Visual display of group of
nodes and their relationships in the

8
GOVERNMENT - The struggle between
hierarchies, markets and networks as modes of
VII. REFERENCES governance in contemporary government.
Retrieved from:
1. Danah M. Boyd, Nicole B. Ellison (2007 Oct). https://www.utwente.nl/bms/mco/bestanden/thepe
Social Network Sites: Definition, History and rspectiveofnetworkgovernment.pdf
Scholarship [Online]. Retrieved from: 12. Sumitkumar Kanoje, Varsha Powar, and
http://onlinelibrary.wiley.com/doi/10.1111/j.1083- Debajyoti Mukhopadhyay (2015). Using
6101.2007.00393.x/full MongoDB for Social Networking Website
2. Pratt CA, Arteaga S, Loria C (2014). Forging a [Online]. Retrieved from:
future of better cardiovascular health: addressing https://arxiv.org/pdf/1503.06548.pdf
childhood obesity. 13. Sebastian Bergmann (2014). Design and
3. Albert-Lszl Barabsi, Zoltn Dezso, Erzsbet Implementation of Workflow Engine [Online].
Ravasz, Soon-Hyung Yook and Zoltn Oltvai Retrieved from: https://sebastian-
(2002, Nov 25). Scale-free and hierarchical bergmann.de/publications/bergmann-
structures in complex networks [Online]. WorkflowEngine-DiplomaThesis.pdf
Retrieved from: 14. Walter W. Powell, Douglas R. White, Kenneth W.
https://www3.nd.edu/~networks/Publication%20C Koput, and Jason Owen-Smith (2005). Network
ategories/03%20Journal%20Articles/Physics/Scal Dynamics and Field Evolution: The Growth of
efree-Hierarchical_Sitges%20Proceedings- Interorganizational Collaboration in the Life
Complex%20Networks,%20.pdf Sciences [Online]. Retrieved from: http://www-
4. F. Toriumi, I. Okada , H. Yamamoto , H. Suwa , personal.umich.edu/~jdos/pdfs/AJS.pdf
K. Izumi and Y. Hashimoto (2011). Classification 15. A B M Moniruzzaman, and Syed Akhter Hossain
of Social Network Sites based on Network (2013). NoSQL Database: New Era of Databases
Indexes and Communication Patterns. Retrieved for Big data Analytics - Classification,
from: Characteristics and Comparison [Online].
http://users.cecs.anu.edu.au/~sguo/swm2011_sub Retrieved from:
mission_10.pdf http://arxiv.org/pdf/1307.0191v1.pdf
5. David F. Nettleton (2015 Oct). A personalized 16. Aravind Mohan, Mahdi Ebrahimi, Shiyong Lu,
privacy tool for online social network users + and Alexander Kotov. A NoSQL Data Model For
Generator for synthetic online social network Scalable Big Data Workflow Execution [Online].
graph data. Retrieved from: Retrieved from:
https://www.researchgate.net/publication/2849006 http://arxiv.org/pdf/1307.0191v1.pdf
14_A_personalized_privacy_tool_for_online_soci 17. Konglong Tang, Yong Wang, Hao Li, Yanxiu
al_network_users_Generator_for_synthetic_onlin Sheng, Xi Wang, and Zhiqiang Wei (2013).
e_social_network_graph_data Design and Implementation of Push Notification
6. Riitta Toivonen, Jukka-Pekka Onnela, Jari System Based on the MQTT Protocol [Online].
Saramaki, Jorkki Hyvonen, Kimmo Kaski Retrieved from: http://www.atlantis-
(2006, Feb 21). "A Model for Social Networks - press.com/php/download_paper.php?id=9566
Harvard University." [Online]. Retrieved from: 18. Wei Tan, M. Brian Blake, Iman Saleh, and
https://www.hks.harvard.edu/netgov/files/fellows/ Schahram Dustdar (2013, Oct). Social-Network-
onnela_j/papers/A7.pdf Sourced Big Data Analytics [Online]. Retrieved
7. Lobna Sameer Hassan, Hany Abdelghaffar (2015, from:
Jun). The use of Social Network in Enhancing e- http://s3.amazonaws.com/academia.edu.document
Rulemaking. Retrieved from: s/32663735/Social-Network-
https://www.researchgate.net/publication/2805547 Sourced_Big_Data_Analytics.pdf?AWSAccessKe
65_The_use_of_Social_Network_in_Enhancing_e yId=AKIAJ56TQJRTWSMTNPEA&Expires=14
-Rulemaking 75214070&Signature=59%2Bugk8n0FJjRqrrGU
8. Sean Whitsitt, Abishek Gopalan, Sangman Cho, KgFgPOpbI%3D&response-content-
Jonathan Sprinkle, Srinivasan Ramasubramanian, disposition=inline%3B%20filename%3DSocial-
Liana Suantak, Jerzy Rozenblit (2012 April). On Network-Sourced_Big_Data_Analytic.pdf
the Extraction and Analysis of a Social Network 19. Michael Beye, Arjan Jeckmans , Zekeriya Erkin,
with Partial Organizational Observation [Online]. Pieter Hartel, Reginald Lagendijk, and Qiang
Retrieved from: Tang (2012). Privacy in Online Social Networks
https://www.researchgate.net/publication/2540314 [Online]. Retrieved from:
99_On_the_Extraction_and_Analysis_of_a_Socia http://doc.utwente.nl/81270/1/Beye12privacy.pdf
l_Network_with_Partial_Organizational_Observat 20. Mason A. Carpenter, Mingxiang Li, and Han
ion Jiang (2012). Social Network Research in
9. Kim, Myunghwan, and Jure Leskovec. Modeling Organizational Contexts: A Systematic Review of
Social Networks with Node Attributes Using the Methodological Issues and Choices [Online].
Multiplicative Attribute Graph Model [Online]. Retrieved from:
Retrieved from: http://jom.sagepub.com/content/38/4/1328.full.pdf
https://cs.stanford.edu/people/jure/pubs/magfit- +html
uai11.pdf 21. Amy Garbrick and Roy Clariana (2012, Apr). The
10. Wei Liu, Matteo Pellegrini, and Xiaofan Wang influence of email notifications in asynchronous
(2014, June 27). Detect Communities Based on discussion on interaction patterns using social
Network Topology [Online]. Retrieved from: network analysis [Online]. Retrieved from:
http://www.nature.com/articles/srep05739 https://www.learntechlib.org/p/15091
11. Jan van Dijk Anneleen, and Winters-van Beek 22. Alan Mislove, Massimiliano Marcon, Krishna P.
(2008). THE PERSPECTIVE OF NETWORK Gummadi, Peter Druschel, and Bobby

9
Bhattacharjee (2007). Measurement and Analysis
of Online Social Networks [Online]. Retrieved
from:
http://conferences.sigcomm.org/imc/2007/papers/i
mc170.pdf
23. Sai Praveen Anumokonda (2014). Asynchronous
Communication in SOA [Online]. Retrieved from:
https://saipraveenblog.wordpress.com/2014/12/08
/asynchronous-communication-in-soamsa/
24. Alessandra Sala, Lili Cao, Christo Wilson, Robert
Zablit, Haitao Zheng and Ben Y. Zhao (2010).
Measurement-calibrated Graph Models for Social
Network Experiments. Retrieved from:
https://www.cs.ucsb.edu/~ravenben/publications/p
df/socialmodel-www10.pdf
25. Sriram Krishnan, Patrick Wagstrom, Gregor von
Laszewski1. GSFL: A Workflow Framework for
Grid Services. Retrieved from:
ftp://140.221.6.23/pub/tech_reports/reports/P980.
pdf
26. Ciro Cattuto, Andre Panisson, Marco Quaggiotto,
Alex Averbuch (2013). Time-varying Social
Networks in a Graph Database. Retrieved from:
http://event.cwi.nl/grades2013/11-averbuch.pdf
27. Rania Soussi, Marie-Aude Aufaure, Hajer
Baazaoui (Jun 3, 2010). Towards Social Network
Extraction Using a Graph Database. Retrieved
from:
http://ieeexplore.ieee.org/document/5477149/
28. Miyuru Dayarathna, Toyotaro Suzumura (2014).
Towards Scalable Distributed Graph Database
Engine for Hybrid Clouds. Retrieved from:
http://datasys.cs.iit.edu/events/DataCloud2014/Da
taCloud2014-paper-1-fullpaper-Dayarathna.pdf
29. Augusto Morales, Tomas Robles, Ramon Alcarria
and Edwin Cedeo (August 20, 2013). On the
Support of Scientific Workflows over Pub/Sub
Brokers. Retrieved from:
http://www.mdpi.com/1424-8220/13/8/10954/htm
30. Katherine Faust. Comparing Social Networks:
Size, Density, and Local Structure (2006).
Retrieved from: http://www.stat-
d.si/mz/mz3.1/faust.pdf
31. Enterprise Social Network (ESN) Vendor
Comparison (2013). Retrieved from:
http://www.dialogueconsulting.com.au/documents
/2013/05/report-enterprise-social-networking-
platform-vendor-comparison.pdf
32. Andry Alamsyah. Big Data Analytics: A Social
Network approach (2015). Retrieved from:
https://www.slideshare.net/andrybrewok/big-data-
analytics-a-social-network-approach

10

Das könnte Ihnen auch gefallen