Follow

Off Premise vs On Premise

BlazeMeter is a commercial, self-service load testing platform-as-a-service (PaaS), which is fully compatible with open-source Apache JMeter.

Testing with BlazeMeter

Target Audience:

Open to all who wish to perform load testing on their website/application.

Prerequisite:

  • Blazemeter Account
  • Application accessible outside your company firewall for testing

Process:

BlazeMeter creates instances on a public cloud using Amazon EC2 which hold the JMeter setup and which will create load on your hosted app. The cloud instances record the app's performance, and then sends these reports back to BlazeMeter in real-time for your viewing pleasure.

The load is generated by a distributed JMeter architecture. A JMeter console is used to control the test. You need to choose the number of JMeterEngine(s) that will participate in the test.

Each JMeterEngine will simulate the number of threads/virtual users specified in the script you provide. For example, 20 engines and a script with 100 threads will generate 2,000 simultaneous users traffic.

Imagine a cluster of up to 100 dedicated JMeter engines, all preconfigured, available in 1 of 11 geographical locations and ready to run 24/7 with no setup at all. No more than 5 minutes is needed to set up and run a test. 

On Premises Testing with BlazeMeter

Target Audience:

For developers who need to test apps behind a firewall, the off-premise solution won't work. They need On-Premise testing.

Prerequisites:

  • BlazeMeter Account
  • Port 80 and 433 open for outgoing connections.
  • Constant connectivity to the internet.
  • On Premises Server should run on Ubuntu or CentOS. They must have a dual-core processor, at least 4 GB RAM, and at least 10 GB hard-disk space free.

Process:

With our new service, no incoming requests need to be made. Simply install our agent on your on-premise servers. This agent will give your servers a "heartbeat" -- by sending outgoing requests to BlazeMeter to check if any tests have been started, and if so, BlazeMeter will respond with instructions for these servers.

They (your warm blooded servers) will then create the load for your DevApp, while sending data back to BlazeMeter so you have full access to our real-time reporting.

Recommendations

We recommend running up to 1000 threads/virtual users per one JMeter Engine if you are using HTTP/S protocol.

The number of users that can be simulated depends on a few other factors.

The number of supported threads depends on your script intensity and available resources. The more intense your script – the fewer threads a single JMeterEngine can support.

For example, a test with 10 JMeterEngine(s), each running 300 threads, simulates a total of 3,000 users.

Please refer to this article for more information on threads per engine/engine per console calibration.

​​Return to the 'On Premise Cluster' article.

 

Want to learn more about running on-premise load tests? Watch our on-demand recording, How to Make JMeter Highly Scalable and More Collaborative With BlazeMeter

Have more questions? Submit a request

0 Comments

Please sign in to leave a comment.