Our goal for each and every client is a stable and secure release. Whether a game, a big data platform, or a website, we want you to be happy and confident in the knowledge that your product will launch without vulnerabilities or stability issues.   For the stability side of our work, we are talking about Load Testing. For this service, we use our proprietary headless technology; Cyrex Swarm. With this solution, we comfortably offer comprehensive performance and scalability testing. So, once we’re done, you will know exactly where and when your system struggles and how to fix it.   Let’s run through the types of load testing to understand what they do and what you might need for your project when seeking efficient scalability.  

Stress Test

A stress test is one of the most common load tests. You start with the traffic quite low and gradually increase the load over time. The goal of this test is to see where your system’s breaking point is. There’s no set duration or user count, just a broad examination of where the system’s limit is. This is good for clients to understand where their upper limit is, whether it’s enough or not for their purpose.   Discovering the system’s limit can inform you of any number of reasons for the crash. For example, it could help our team discover a bottleneck. Once that is discovered, we trace backwards to see which part of your program is causing it. It can be as simple as a setting not adjusting to the rising users or something more complicated such as a hard limit on what it can handle. Whatever number we hit during this stress test is then weighed by the client for their need. If it needs to be higher, we advise on how best to adjust and grow.  

Spike Testing

A spike test simulates high user traffic in a short period of time. This is good for big event launches or updates. Things like NFT drops or the launch day of a game. The shorter timeframe means an exponential increase in intensity. Your system might be able to handle a gradual increase to 20,000 over time but it also might need to handle 5,000 in a quarter of the time.   This is a great example of how scalability isn’t just about size. It’s also about efficiency in handling requests and new users in any timeframe.

 

Peak Testing

Peak testing involves matching the traffic expected for the platform’s busiest period. For accountancy platforms, the end of the month is always busy. For gaming, servers usually have their busiest period between 5 p.m. and 9 p.m.   The difference in peak testing is the flow of normal users to higher than average. We’re testing to see if the system can handle a jump in traffic and then it will adapt when traffic lowers to its usual numbers. It’s also slower overall than a spike test. This is important because scalability goes both ways, your system must not allocate resources when they’re not needed.  

Soak Test

Finally, the soak test. The longest by a wide margin, this test occurs over a 12 to 24 hour period with numbers equalling slightly above normal traffic for the platform.   If you have regular backups every 6 or 12 hours, or have a regular process like game leader boards which might update every hour or two, you’re ensuring the process works normally over a long period of time with higher than normal traffic. These types of processes need to work without any issues and need to not interrupt or cause any further problems for the usual play or operation.   This test ensures your integral and occasional updates work without fail regardless of higher traffic. It also helps to look for memory leaks in gaming, if the cache only resets at the end of a session. Or in a protocol, if the number increases with each packet with a set limit, you’ll eventually get a packet overflow. We support each of these types of testing and have the expertise to deliver solutions for any issues that might occur. With technology like Cyrex Swarm, we’re capable of delivering up to half a million users easily! Get in touch today and ensure your system is stable and scalable.