The presentation on Performance Testing of Big Data Application was done during #ATAGTR2017, one of the largest global testing conference. All copyright belongs to the author.
Author and presenter : Harpreet Kaur Kahai
2. Agile Testing Alliance Global Testing Retreat 2017
Importance of Big Data:
Source: https://www.slideshare.net/OECD-DAF/big-data-bringing-competition-policy-to-the-digital-era-background-note-oecd-competition-division-november-2016-oecd-discussion
3. Agile Testing Alliance Global Testing Retreat 2017
Challenges with Performance Testing of Big Data Application
Performance testing activity plays a very important role in achieving the desired
performance of application meeting the SLA by evaluating the data capacity size without
defeating the goal of building Big Data System
Common Challenges are as follows:
• When there are so many parameters that are modified it is very difficult to test a big
data system by extrapolating from an undersized test system
• Creating Data is a key issue in big data application benchmarking as creating TB or PB
scale real big data sets in undersized test system can be very expensive and time
consuming
• Backup and Restore of Big Data is not possible while Performance Testing Big Data UI
Application like traditional application while dealing with Terabytes and petabytes of
unstructured & semi structured data
• Identifying Performance Bottlenecks, sharing performance test execution results and
followed by performance tuning
4. Agile Testing Alliance Global Testing Retreat 2017
Performance Testing Approach
Performance testing process begins with understanding the architecture of Application and
various components that are integrated with Big Data Systems
5. Agile Testing Alliance Global Testing Retreat 2017
Performance Testing Approach
A big data application makes use of the following set of technologies
• Map Reduce frameworks
• NoSQL databases
• Message Queue
• Search components like Elasticsearch
11. Agile Testing Alliance Global Testing Retreat 2017
Performance Testing Approach
Gems is a GUI utility for Enterprise Message Service (EMS). It is tool used for
monitoring. Messages may be sent or received, message queues contents can be
inspected. Selectors and filters can be applied. Accordingly, if depth of queue
increases, count can be viewed and messages can be purged.
13. Agile Testing Alliance Global Testing Retreat 2017
Solution/Features of a Framework
Bottleneck are identified using various tools like App Dynamics, Grafana, Gems, Analysing DB queries and Checking
Logs. Following are the areas generally taken up by Development Team for performance optimization:
o Server Compression
o JavaScript Cache
o CSS Cache
o Hibernate/SQL Tuning/Database Tuning
14. Agile Testing Alliance Global Testing Retreat 2017
Conclusion
Performance Testing of Big Data Application is a challenging task. Every organizations have to choose the best
solutions according to their needs, to solve their performance testing challenges. Above tools can be used for
running performance tests to identify and resolve bottlenecks
Editor's Notes
Big Data helps organization achieve several objectives. It stores the information in various types and format which is used for decision making like
Understanding customer needs to Forecast demand and utilization
Making relevant offers for continued usage of online services as downtime of application is greatly reduced
Maximize product/service quality by identifying improvement to be made in infrastructure and service delivery
Becoming more innovative and competitive by validating the real-time data
Optimize Workforce planning and operations
Discover new sources of revenue and also improves return on investment
Using right channels for communication with customer
Big Data provides various tools, methods and technologies which are used to capture, store, search and analyze the data to establish new correlations, relationships and trends that were previously unavailable
Performance testing activity plays a very important role in achieving the desired performance of application meeting the SLA by evaluating the data capacity size without defeating the goal of building Big Data System
Common Challenges are as follows:
When there are so many parameters that are modified it is very difficult to test a big data system by extrapolating from an undersized test system
Creating Data is a key issue in big data application benchmarking as creating TB or PB scale real big data sets in undersized test system can be very expensive and time consuming
Backup and Restore of Big Data is not possible while Performance Testing Big Data UI Application like traditional application while dealing with Terabytes and petabytes of unstructured & semi structured data
Identifying Performance Bottlenecks, sharing performance test execution results and followed by performance tuning