Conference Lecture Review - Rosy Upreti

ickybiblegroveInternet και Εφαρμογές Web

3 Νοε 2013 (πριν από 3 χρόνια και 10 μήνες)

61 εμφανίσεις



Upreti 1

Rosy Upreti

April 3
,

2012

Name of the Conference: 28
th

IEEE International Conference on Data Engineering



CONFERENCE
LECTURE

REVIEW

On
April 3, 2012
, I

attended the conference

presentation

on “BestPeer++
: A

Peer
-
to
-
Peer based
Large
-
Scale Data Processing Platform”

host
ed

by Gang Chen. The presentation focused on the
integration of the technologies such as cloud computing, database, and P2P technologies into one
system that contributes towards the development
of scalable, flexible and economical platform
for corporate network applications. The conference also emphasized towards the deliverance

of

data sharing services to participants based on the widely accepted business model.


Chen,
after introducing himself

and the co
-
authors who worked together for the accomplishment
of this paper, started explaining the importance of corporate network in sharing information
among the participants companies and collaborating in a certain industry sector where companies
typi
cally share their common interest. He also pointed out the fact that this type of network helps
the companies in reducing the cost of the operation and increases the revenues. However, Chen
also discussed about the challenges that the company had to face w
hile doing the inter
-
company
data
sharing and

processing. He discussed that during this process there is the high risk

in
scalability, performance, throughput and security if the management are not done properly. The
presenter then highlighted the BestPeer
++ system [1], that he and his coworkers have proposed
to deliver the elastic data sharing services for the corporate network applications in the cloud
based on a peer
-
to
-
peer (P2P) data management system. The author also highlighted on the
evaluation that

they have done for a proposed model on Amazon EC2 cloud platform which


Upreti 1

ultimately resulted in the linear scalability for throughput with respect to the number of peer
nodes[1].

The proposed system discussed by the presenter Chen, focused on choosing the r
ight data
sharing platform that enabled the shared data to be visible network wide

and supported the
analytical queries over those data. The service was deployed in the cloud and the corporate
network was formed by simply registering their sites with BestP
eer ++ service provider [1].
Then, BestPeer++ instances were launched in the cloud that finally exported data to those
instances for sharing [1]. Chen addressed that the system adopts the pay
-
as

you

go business
model popularized by cloud computing, which

, according to him reduction of the total cost of
ownership as the companies are not required to buy any software in advance. Presenter Chen in
his conference presentation
also focused that the proposed system was also compared with
HadoopDB, which was th
e recently proposed large scale distributed system with a set of
designed for data sharing applications and it was found that the performance of BestPeer++ was
significantly better than the former system. Chen, in his presentation provided an overview of t
he
proposed system along with the design description for the core components. The query
processing strategies that was

adopted in the system was also
discussed by the presenter

Chen.
Chen finally concluded his speech by presenting the evaluation that was d
one on the
performance and throughput of the proposed system on Amazon Cloud platform [1].


The discussion on pee
r to peer network has

also done by Henskens and Hannaford on their paper
on their paper “A System for Robust Peer
-
to
-
Peer Communication with Dy
namic Protocol
Selection”[2].
The

paper analyzes the dynamic system that is capable of supporting peer to peer
communication in a large array of network designs that ranges from small home networks to
large corporate networks that consist of the firewalls

and proxies. There is the introduction of


Upreti 1

new concept of service that helps in providing peer to peer communications between and among
the clients [2]
. The

paper examines the service as being capable of operating in different
depending upon the support av
ailable from the network surro
unding.
There is also a paper on

Approximate Aggregations in Structured P2P Networks” [3] presented by Dalie Sun and his
friends which discusses sampling based scheme that contributes towards the speeding up the
processing of

the aggregate queries. The paper introduces the query processing strategies for peer
to peer corporate networks by taking the sample from the local data base. The utility of the
synopsis was maximized by adjusting the synopsis based on the histogram [3].
The discussion on
cloud computing requirements has been done in the paper on “Satisfying Cloud Computing
Requirements with Agile Service Networks” by Tamburri and Lago [4]. This paper proposes the
agile research network that satisfies the cloud computing r
equirements in software engineering.
This paper identifies the list of requirements for cloud computing application and sets the stage
for more research on the Application of agile service networks to Cloud
Computing [4]
. The
challenges of the cloud
computing is also highlighted on the paper along with the solution to
overcome it by thorough investigation of the Agile research network
. All these scholars
emphasizes on the importance of the technologies like peer to peer and cloud computing which
reinf
orces the presenter’s topic in the conference presentation to be more efficient and relevant in
the present context.

Overall, I found the presentation to be very interesting. It was a research oriented topic that
emphasized on the development of the more e
fficient corporate network

with the use of peer to
peer technologies and cloud computing
. T
he topic helped me in understanding the different
challenges associated in sharing and processing data
in inter
-
business
environment. The

presentation also helped me

critically analyze the technological aspects associated with the


Upreti 1

database management such as cloud computing, peer to peer technologies etc. Also knowing,
about the query processing strategies and their consequences in data processing was really
exciting.

Attending this
conference

has not only helped me in participating in the conversation
that my discourse community engage in but

has also raised by enthusiasm towards the real
application of my graduate courses in the industrial field.


Reference
:

[1] Gang

Chen, Tainlei Hu, Dawei Jiang, Peng Lu, Kian
-
Lee Tan, Hoang Tam Vo, Sai Wu,

“BestPeer++: A Peer
-
to
-
Peer based Large
-
scale Data Processing Platform”,
IEEE
International
Conference On Data Engineering
,
April 3, 2012
, 1
-
12

[2]
Wa
llis, M, Henskens, F, Hannaford, M, “A System for Robust Peer
-
to
-
Peer Communication
with Dynamic Protocol Selection”,
IEEE Conference on Parallel and Distributed Computing,
Applications and Technologies
,

3
-
6 Dec. 2007, 161
-
162

[3]
Dalie Sun, Sai Wu, Shouxu
Jiang and Jianzhong Li, “Approximate Aggregations in
Structured P2P Networks”,
IEEE Transactions on Knowledge and Data Engineering
, vol 23, no
11, November 2011

[4]
Damian A. Tamburri, Patricia Lago, “Satisfying Cloud Computing Requirements with Agile
Serv
ice Networks”,
IEEE World Congress on Service
,

4
-
9 July
, 2011
, 501
-
506