What is our primary use case?
I have been using it to analyze the performance of various enterprise systems like SAP, web applications, Oracle Labs, and mobile app applications as well. The objective of the performance testing is to assess the system and whether it can withstand an actual load.
How has it helped my organization?
One of the main benefits is that we don't need to worry about load generators. Before we had Loadrunner Cloud, we had 10 to 20 load generators, and we needed to maintain them. There were always upgrades. We would also have to configure the load generators when we were ready to test, and at times there was stuff to clean up. None of that is necessary anymore. We can concentrate most of our time on creating scripts and configuring Loadrunner based on our objectives. It's very easy for us. We don't need to worry about what happens in the backend or how the load is distributed.
A second advantage is the reporting system. Once a test is done, even if it's in continuous integration and continuous delivery, it automatically triggers and produces reports. We can send the reports to multiple dev teams whose developers are expecting the reports.
Another advantage of the solution is seen compared to others we have used, like JMeter, and IBM solutions. LoadRunner has reduced a lot of the time involved in the scripting cycle because of TruClient. We only need to record one time and we can configure it accordingly to create various scenarios. It has reduced scripting time by about 50 to 60 percent.
When we do mobile applications, every build can be tested with LoadRunner Cloud. Once the developer checks into a particular branch, we call it a mod branch, the performance test is already integrated into it. They can trigger it through their continuous pipeline. For mobile development platforms, it's very crucial because notifications and other things need to be tested before we send to prod. Notifications are very load-sensitive because they go to every user—it could be 10,000 or 20,000 users—and we previously had issues with this. Using LoadRunner Developer and continuous integration has helped us.
What is most valuable?
Recently I have found the TruClient tool to be very useful. My team is involved in a lot of performance testing of applications including mobile platforms and different types of web browsers. In those cases, LoadRunner TruClient is very crucial to testing the performance. TruClient is one of the best features, one I use every day to create load scenarios.
The fact that the solution supports multiple protocols such as open source, VuGen, TruWeb, TruClient, and SAP is very important because these protocols help us to concentrate on what is really needed to produce performance tests. If something is not supported, you have to use other tools or find other ways of assimilating loads. For example, when you are trying to create loads for web applications, if it's not TruClient, you need to find and analyze every call: What the HTTP request is, and what the other kinds of requests are that we need to call, and then correlate all the correlation identifiers. All of that is taken care of by TruClient and other protocols. That is actually very beneficial across all the platforms. For example, a SAP application might be tested for mobile. We can use a combination of the SAP protocol and TruClient protocol. The combinations are very helpful for performance testing.
LoadRunner Cloud gives you a lot of options, even for multi-browser or multi-device testing. It has been the main tool that can do everything; complete end-to-end performance testing.
The support for large-scale testing is also an important feature in our operations. We have thousands of users and it provides the best solution. You can have an unlimited number of users, although you would need to pay for them, but that's a different story. In our organization, the maximum number of users is currently about 20,000 to 30,000. It's a one-stop solution. I can configure my load on the cloud environment and have 30,000 virtual users on the cloud. I don't need to create the infrastructure locally and I don't have to maintain it. Everything is taken care of by the solution.
In addition, its ability to run unlimited concurrent tests without worrying about hardware availability helps eliminate hardware dependency. You don't need to have the load generators on your network. You don't need to maintain those systems and you don't need to have that kind of network capability. If you're testing on-prem, but you don't have the network capacity to scale up to millions of users, LoadRunner Cloud enables you to create virtual networks and use the cloud to generate those kinds of loads. You can then analyze what the impact will be to your system when you have millions of users. LoadRunner Cloud is the best way to do that.
What needs improvement?
When it comes to the CI pipeline, there were some limitations initially, but the latest version of LoadRunner is very helpful. They can integrate into the CI/CD pipeline. We are trying to put it into a complete CI/CD pipeline, but there are still some challenges when you try to run it through different protocols. The challenges are around how you can containerize applications. There are some limitations to some protocols, such as desktop. And when it comes to database testing, there are some things that we can't do through CI/CD.
For CI/CD, the previous versions may not be the right ones, but the latest version is definitely a step ahead. We are aiming for 100 percent, but we have achieved around 60 to 70 percent in CI/CD. Still, it's very good to have that capability.
Also, it would be helpful if Loadrunner Cloud had the same kind of enterprise environment where we had multiple models and options while creating the load profile. Not all the options are available in the LoadRunner Cloud. If they could be added, it would be good.
For how long have I used the solution?
I've been using a LoadRunner solution for more than 10 years as part of my performance testing.
What do I think about the scalability of the solution?
The efficiency of the scalability is a 10 out of 10. I have used multiple tools and LoadRunner is the best one in terms of efficiency. When it comes to cost-effectiveness of the scalability, I would give it an eight out of 10. Even though it's cloud, and you can have thousands of users, we are paying in the tens of thousands. It's not so cost-effective for a university like ours. We still have to justify why we need to spend so much money every financial year.
How are customer service and technical support?
I have used Micro Focus technical support a couple of times and they have been very supportive. They're good.
How was the initial setup?
The only problem with setup is that there might be some problems with the firewalls, configuring SSH and other things. We were initially using Micro Focus SiteScope but we had some issues getting all the metrics. With New Relic and Dynatrace, we stopped using SiteScope. Other than that, we didn't have any issues. It's pretty straightforward. You install LoadRunner, configure your virtual users, and create the environment.
Our initial deployment of the on-prem solution took a week to get everything done, including setting up the firewall, configuration files, and the protocols. The migration to Loadrunner Cloud is nothing. You can start triggering whatever code you had on-prem in the cloud. There really isn't any migration involved. It's pretty straightforward.
What was our ROI?
We have definitely seen return on our investment with Loadrunner Cloud. As I noted, we used to have many load generators and they are no longer used. That is saving us about $30,000. And Loadrunner costs 50 percent of what it used to cost us to run the same test in the cloud. We don't need a Windows Server license anymore or networking capabilities specifically for testing. Those are the kinds of savings we have seen from moving to the cloud.
Also, in the past, we used to write code. But with TruClient, while you need a performance tester, you don't need a programmer to write scripting. If you know the system, and if you know the objectives of performance testing, you can do the performance tests. No programming skills are needed. That also gives us leverage. We can use someone with performance testing capabilities, even though he might not be the best programmer. That has also reduced our costs by $10,000 a year.
Which other solutions did I evaluate?
I have extensively used JMeter as a performance testing tool. JMeter is free and also enables codeless scripting. Even without coding you can try running scripts. But where there is no comparison between JMeter and LoadRunner is when it comes to distributing load. LoadRunner stands out. With JMeter, it's very difficult to distribute the load.
When it comes to creating reports, Loadrunner is best. You will spend most of your time analyzing what's happened with the test, analyzing bottlenecks and pain points with the performance parameters. But in JMeter, you have to manually collect everything: collate the results and produce the reports. Then you need to do a detailed analysis to find the bottlenecks and resource patterns. It is very difficult, but it's free. If you have the skill set and the time, you can use JMeter. But if you are time-constrained, and you want to actually concentrate on performance testing, use LoadRunner.
LoadRunner Cloud provides application performance and management tools to an extent, but not to the extent of New Relic or Splunk. We predominantly use New Relic to monitor application performance and in some cases we use Dynatrace as well. But LoadRunner Cloud doesn't have complete application performance monitoring metrics.
What other advice do I have?
LoadRunner's Developer integration enables developers to script and run tests without leaving the developer ecosystem. It gives a complete IDE where you can develop the code and add your script. For example, if you are using a Java platform to code, and it has all the libraries and the IDE, you can integrate your load testing into your development. For us, because we don't have a single development cycle or ecosystem—we constantly move to various methods—that's where the IDE has limitations. It supports certain languages but it doesn't support everything. If we are using Go, for example, we might need to add certain libraries, so that's where it isn't helpful. But if you are purely on Java or core platforms, it will definitely help.
The Developer integration enables developers to add performance testing to their day-to-day tasks, but shift-left depends on your company's development strategy. If your whole culture supports shift-left, so that your quality assurance is embedded from the start of the development cycle, then shift-left works out. But in our case, we use a lot of packaged applications like Workday or SAP where we don't have much opportunity to work with their code. We do a lot of configurations and integrations. In that case, shift-left doesn't work as well. But whether you follow shift-left or shift-right, Loadrunner Cloud works.
For non-functional testing, LoadRunner is the best tool. I can recommend it to other people, to create specific tests from the smallest load to the highest level of load. I don't think any other commercial tool has that capability to create load performance testing. There is no other tool that gives this kind of experience for a load-testing professional. From end-to-end, starting from creating the load scenarios, to running them, and then reporting, LoadRunner is the best tool. You save a lot of time and, with LoadRunner Cloud, you are saving a lot of money. Go for it.