The goal of benchmarking is to understand the performance capabilities and limitations of solution built using IBM SFG. Increase in the number of partners/transactions will lead to increase in file volumes as a result benchmark results will have a short shelf-life - this is a domino effect. So, benchmarks needs to be determined on a regular basis to estimate the capacity requirements for the anticipated volumes. Considering this regular activity which is required for a high performance IBM SFG solution - it is important to setup a reusable benchmarking methodology to save time and cost in executing the benchmark tests.
Remember, there are many ways to improve performance of IBM SFG solution – so, nothing to worry by looking at performance numbers. IBM Software Services may be engaged for tuning, improving the performance and optimizing the solution design.
Now, for a very simple “Inbound FTP test case” scenario - let’s see how to artfully capture performance numbers while benchmarking IBM SFG solution using Apache JMeter. Apache JMeter is an open source testing tool which has ability to load and measure IBM SFG solution performance.
- IBM Sterling File Gateway v 220.127.116.11
- Apache JMeter v 3.2
- Apache JMeter Plugin – PerfMon, Graphs Generator, Server Agent
2.) SFG setup
2.1) Configure FTP Server Adapter
2.2) Create a producer, consumer partner
2.3) Create a static routing channel template
2.4) Create a routing channel
3.) JMeter setup
3.1) Create a test plan as below
3.2) Create a thread group as below
3.3) Add and configure FTP Request Send sampler as below
3.4) Add and configure CSV Data Set Config
3.5) Add and configure following listeners
4.) View results (after running the test)
4.1) Summary Report
4.2) Bytes Throughput Over Time
4.3) Hits per second
4.4) Latencies Over Time
4.5) Transactions Per Second
4.6) Response Times Over Time
4.7) Health of SFG Server
Now you know the basics of measuring the FTP put performance. Your initial performance numbers can be improved greatly by tuning IBM SFG solution and re-benchmarking to validate the tuning parameters.
We will go deeper into Advanced benchmarking / testing / measuring techniques in future blog posts.
Note: For heavy test scenarios run JMeter in non GUI/command line mode or as recommended by Apache JMeter. It is not mandatory to use Apache JMeter for benchmarking IBM SFG solution - there are many testing tools available in the market which can be chosen depending on the convenience and enterprise’s testing methodologies.
Please do not cite, circulate, or copy without permission of the author. The views expressed are solely my own and this article do not express the views or opinions of my employer. Neither myself nor IBM recommend any approach or tools mentioned in the article. I shall not be accountable for and does not accept any liability or responsibility whatsoever for any inaccuracies, omissions, errors, misleading information, loss or damage howsoever arising (including without limitation, incidental, punitive, exemplary, special or consequential damages, loss of profit or damages resulting from lost data or business interruption), be it by negligence or otherwise, or expense resulting from:-
- Using the information and contents of this article whether with or without authorization;
- Relying on the information and contents of this article, whether downloaded or not;
- The performance, operation, functionality, non performance, unavailability, inaccessibility, or corruption of this article;
- Refer vendor specific documentation for any kind of latest information;