Page 1 of 5
Journal for Studies in Management and Planning
Available at
http://edupediapublications.org/journals/index.php/JSMaP/
ISSN: 2395-0463
Volume 02 Issue 08
AUGUST 2016
Available online: http://edupediapublications.org/journals/index.php/JSMaP/ P a g e | 510
Standardization of Stress Test Protocol for successful implementation of
Agricultural Applications – A Study at ICRISAT
VV Sumanthkumar1
1Scientist from International Crops Research Institute for Semi-Arid Tropics, Patancheru,
Hyderabad.
Introduction
The new Sustainable Development Goals have once again underlined the importance of
agriculture development if want to achieve global food & nutrition security and end poverty. On
one side the agriculture is effected with myriad of challenges like climate change which is
effecting agricultural production and productivity. To mitigate the major challenges and making
the farmer resilient to these negative forces, provision of knowledge and services directly to the
farmers is one of the important aspect which certainly cannot be ignored. The advances in
computing hardware and ICT infrastructure really allowing to implement knowledge as well as
services directly to the farmers. The current ICT infrastructure in rural village is improving with
advancement of National Optic Fibre Network (NOFN). NOFN which is popularly called as
Bharat Broadband to connecting all revenue villages (Gram Panchayats) of India with 100 Mbps
band width. Infrastructure in villages is improving but still there are challenges with respect to
network in rural areas. It is ideal to keep optimal resources at server side so that farmer gets
response quickly which keeps him attached and attracted to the application and thus the adoption
of application increases. ICRISAT is in the process of designing many spatial applications to
different clients. As a process of building successful applications, there is need to standardize
and optimize computing resources at server side. An attempt has been made towards this
direction by developing a protocol for stress test which can be used to optimize the resources
while developing agricultural applications for masses.
Discussion:
The paper especially discusses about following key initiatives and learning from it: When any
organization builds huge application which requires a data center typically follows setup so as to
serve the millions of small holder farmers. The below is the generic layout followed.
Page 2 of 5
Journal for Studies in Management and Planning
Available at
http://edupediapublications.org/journals/index.php/JSMaP/
ISSN: 2395-0463
Volume 02 Issue 08
AUGUST 2016
Available online: http://edupediapublications.org/journals/index.php/JSMaP/ P a g e | 511
But there are instances where requirements may vary based on the data stored, processed and
accessed by the end user. One important factor is whether the application stores spatial data or
non-spatial data. GIS applications which processes spatial data generally requires more resources
when compared to MIS applications which deals with non-spatial data. The protocol is
developed to finalize the ideal resources required for a GIS application which hosts spatial data.
Generally, the GIS project intends to support in the planning process through characterization of
natural resources, establishment of a digital library thus enabling all spatial and non-spatial data
related to Micro watershed and agricultural fields to all stakeholders through role based access
over web. Land information being critical for above, it is to proposed to develop a dynamic state- of-art GIS based Land Resource Inventory (LRI) Geoportal which can ingest, organize, manage
and disseminate data from multiple sources (spatial and non-spatial) for enhancing decision
support to the stakeholders including planning and sustainable development of land resources for
various applications. The proposed system shall serve as a common operational platform
providing actionable intelligence and situational awareness to all its stakeholders. The proposed
system is also expected to meet the national / state level objectives in regard to sustainable
development of agriculture, disaster planning and recovery, infrastructure planning and natural
resources assessment and allocation, rural development, education, etc. As this application
involves intensive data transmission from central database to end user, it is essential to make sure
that data retrieved quickly without delay. This is essential for making the application adopted by
different stakeholders of the project successfully.
Page 3 of 5
Journal for Studies in Management and Planning
Available at
http://edupediapublications.org/journals/index.php/JSMaP/
ISSN: 2395-0463
Volume 02 Issue 08
AUGUST 2016
Available online: http://edupediapublications.org/journals/index.php/JSMaP/ P a g e | 512
To make sure that sufficient computing resources are available at cloud, we have conducted a
stress test with the following server configurations for different servers like Open Source Geo
Server and Server that generates Load.
Description Processor RAM Storage OS /
Software
Open Source
Geoserver
2 Xeon
(4 Cores)
2.5 GHz
16 GB 1.9 TB JBOSS
Application
Server
Load Generator
Server
Dual Xeon
3.16 GHz
4 GB 500 GB JMeter Load
Generator
There are so many load generators are available like Tsung and JMeter are available. The
Apache JMeterTM application is open source software, a pure Java application designed to load
test functional behavior and measure performance. JMeter has been widely adopted and is now a
popular open-source alternative to proprietary solutions like Silk Performer and LoadRunner. It
is a cross platform and can run on any platform with Java.
Key features of JMeter
Load and performance test various server/protocol types including http/https,
SOAP/RESTful Webservices, FTP, DB via JDBC, LDAP, Mail, MongoDB, Native
commands/shell scripts and TCP.
Cross-platform – can run on any OS with Java
Multithreading framework – allows concurrent sampling
Faster test plan building and debugging – can be integrated with the test plan; ability to
create a functional test plan
User-friendly UI
Easy integration with major Build and Continuous Integration systems – this makes
Jmeter a well-fitted part of whole Software Development Life Cycle
Methodology and Bechmark:
The JMeter Load Runner controller was configured to increase by 25 users every 90 seconds
until there were 500 users, and then run at that user load for 10 minutes. The Server instance
