Earlier this week we spoke to Tim Koopmans who is the founder of Flood.IO. Flood IO is a distributed load testing platform that leverages cloud infrastructure and open source tools to simulate really large performance testing workloads. Flood.IO provides a scalable platform for load generation with support for Jmeter and Gatling.
Tims responses below will help you understand more about Flood.IO, the platforms capabilities from a Performance Testing standpoint, including how you might be able to save your organization a tidy sum by using Open Source tools combined with Flood.IO’s scalable performance testing platform.
What is Flood.io all about – Flood IO is a distributed load testing platform for everyone. We combine great open source tools like JMeter, Gatling and Elasticsearch, deploy via AWS, and deliver a cost effective SaaS load testing platform. Many of our users are already familiar with open source tools like JMeter and Gatling. We just make it easy for them to scale out with massive concurrency and throughput. We also provide simple to consume and easy to share reports. Basically we take the hassle out of scaling, managing and reporting from your own distributed performance test infrastructure.
Tell us more about yourself and what you do at Flood.io – I’m the CTO of flood.io which means I’m responsible for the technical direction of the product. On a day to day basis I do much of the development and technical support.
How do you segment the various different product offerings at Flood.IO – The product offering is simple, we provide a scalable platform (Grids) for load generation (Floods). Users can choose between Apache JMeter or Gatling to run their Performance Tests or Floods in our terminology. You upload your test plan (or use our API), select which Grid you want to run your flood from and hit start. You then see results from your Flood in near real time.
Users can share results with others via a simple url: https://flood.io/d384673f64e3a3
A Grid is just a collection of semi autonomous nodes. Each node can operate on its on or in a cluster within a geographic region. We support 8 regions across the globe. When you want to scale out for bigger Floods you simply start more Grids. We operate on a monthly subscription and registration is free. We have different levels of subscription which influences on demand costs and data retention of results. Grids are invoiced per node per hour on demand.
Users can change their plan at any time so we’re quite flexible for those that just want to run a single test effort, or for Professional and Business users who are looking to establish and maintain a performance test capability.
Who are Flood.IO’s competitors and how does Flood.IO standout against the competition – Obvious competitors are BlazeMeter and similar. The key things that make us unique:
- We use readily available open source software and offer options like JMeter or Gatling, depending on your requirements.
- We implement a shared-nothing infrastructure where each Grid node is independent and self-sufficient, with no single point of contention. This lets us scale out massively unlike other solutions which rely on centralised ‘controllers’.
- We let Professional and Business users host Grid nodes on their own AWS accounts, passing on infrastructure costs and thereby presenting significant savings.
- We put no limits on how you test. Run as many users as you like and test as often as you wish. No crazy licensing or time / quantity based constraints.
What’s behind your love for Jmeter and Gatling – JMeter has been around for a long time, since 1998. It’s the swiss army knife of open source performance testing. It has wide protocol support and is actively supported by dedicated and talented developers. Its more recent releases have become very high performing. For example some of our own benchmarking demonstrates good performance from a single Grid node with 10K concurrent users: https://flood.io/blog/11-benchmarking-jmeter-and-gatling
Gatling on the other hand is relatively newer. Part of its lustre for us was its use of Akka, which uses event driven asynchronous messages and leads to some massive concurrency from a single JVM. It also has a Scala based DSL. We’ve found Gatling is becoming more popular amongst our ‘developer’ type customers, although that’s a pretty loose generalization.
Do you see yourself supporting any other open source tools in the future –
Absolutely, for a long time we’ve been evaluating diffe
rent WebDriver derivatives, although we’re concious of not putting too many holes in the ozone layer. We’re really after load testing tools that can map potentially thousands of virtual users per core.
Will Flood.IO venture into the Functional Testing space –
I don’t think so. We see Flood as a platform for automating processes which need big concurrency or wide horizontal scale. We also feel that the functional test automation space is well served by products like saucelabs.com
What is the nature of integration between New Relic and Flood.IO – Essentially we let our Frequent, Professional and Business users publish results from their Floods into their New Relic dashboards in real time, so they can view things at a global scale, and readily correlate results from Floods with their application / server metrics collected by New Relic. We offer similar integration to Flowdock and will continue to expand our 3rd party integrations.
What kind of customers are potential Flood.Io users – Our mantra is “load testing for everyone” so potential users is anyone from those dipping their toe into the cold waters of performance for the first time, or the ‘icebergers’ who are veterans of the long distance event, and just need a suitable platform to swim with ..
Do you offer Consulting services to support your customers who might need help with scripting/execution/analysis –
All of our subscriptions are backed by online support, this is a strength for us and our Australian customers. Individually you can also find me on airpair.com
where I offer 1:1 assistance. We’re also well connected with test agencies and consultancies who have resources to support this kind of thing.
Tell us more about some of the projects and customers who use Flood.IO – You can see some of the companies that use Flood IO on our front page. Ninefold use us to help benchmark their hosting infrastructure using JMeter and our Ruby DSL. Zeebox test their mobile social platform with massive concurrency using Gatling. We also get a lot of individuals or small to medium sized businesses as well as test agencies running anything from simple through to sophisticated performance test suites.
How does one go about learning to use Jmeter and Gatling for Performance Testing – Your best bet for either tool is to download the source or binaries and launch into it.
With Gatiling or JMeter the user and developer community is very supportive, so I’d also encourage you to ask questions on the relevant forums, for almost immediate feedback. Here are some relevant forums –
Do you have any advice for budding Performance Engineers and Performance Testers – It’s great if you get the chance to actively configure, build or develop (code) on the target system that you are testing. These days more than ever, this is possible. Ask questions, don’t be afraid to experiment and follow a scientific method.
You can read more about Flood.IO and check out their services at – Flood.io. You can reach Tim Koopmans here – LinkedIn.