Okay so ORIGINALLY the idea was to do a YouTube video to explain the interaction and relationship between the “Test Duration” of a CyberFlood test and the “Load Specification”.

However, in order to best explain “load specification” though it is important to realize that depending on what you are testing, the “load” type you configure in the test will change. For example, for a “throughput” test the load would likely be trying to hit a certain “bandwidth” value. Similarly for maximum number of open connections the “load” you are likely trying to hit is exactly that “# of open connections”. So I knew I also needed to also go into some of the varying CyberFlood “starter test templates”.

If you want to go right to the YouTube and not read the “overview” below link feel free to click here. I do suggest coming back here though still as I will summarize some stuff here I did not there. And yes… the YouTube video has ALL the very detailed and specific timestamps in the YouTube notes so you can just go to the portion (once you go to the YouTube) that you want to see and only that portion.

Let’s hit the “big ticket” items here.

Duration and Load Specification Are Related:

The “Duration” of a CyberFlood test is the “how long the test will totally run”. So for example, the duration of the test below is 6 minutes and 30 seconds.

The “Load Specification” for a CyberFlood test is “what load to try AND at basically what to try to do during every second of the test duration”.

From the load specification above we can see that at the “max height” of the test we are supposed to try to hit 6 Gbps of bandwidth. Notice, though, how there is idle periods at the beginning and end and also a ramp up and ramp down. Let’s take a closer look at this.

The “5” common phases:

There are 5 “common phases” with layer 4 thru 7 traffic generators.
For many people who aren’t used to Layer 4 thru 7 traffic generators the above typical phases in these types of traffic generators is really quite confusing. I promise to dive into and explain in a future YouTube or blog.

For now I will at least just give you one example for when you might use a ramp up time. Going to keep this super easy with silly low math. lol.

Say you have a device you are testing that can have a total of 500 HTTP connections open simultaneously. You WANT to test “max open” and have your steady state testing be “500 open HTTP connections. However the device can only do 50 new connections every second. So you can’t just start with 500 new connections as soon as the test starts. You have to “ramp up”.

When You Change the Duration it ONLY Impacts Steady State:

When you change the duration of a test, that change – be it an increase or a decrease in time – ONLY happens to ONE of the 5 phases. That phase is the “steady state phase”.

Let’s look at this with the above 6 minute and 30 second test, with a 3 minute steady state, as the “case study” of what would happen:

1. ADD 3 minutes to Duration: If you changed the duration from 6 minutes and 30 seconds to 9 minutes and 30 second …. this 3 minutes will get added to the steady state… making it a 6 minute steady state.
2. Subtract 3 minutes from Duration: If you changed the duration from 6 minutes and 30 seconds to 3 minutes and 30 seconds…. this 3 minutes will get taken away from the steady state… making it a 0 second (non existent) steady state.
3. Attempt to Subtract MORE than 3 minutes from the Duration of this Test: If you attempted to change the duration of the above test to MORE than 3 minutes, you will get an error message that this is NOT possible. This is because, again, the only phase that is directly impacted when a duration change is made, is the steady state phase. And the steady state phase for this test is 3 minutes. You cannot, then, decrease the duration of the test more than the duration of the steady state.

The CyberFlood “Starter Test Templates” and their varying default values:

In my job I use both CyberFlood (Spirent) and BreakingPoint (IXIA/Keysight), two companies that have been doing traffic generation for years and years. Why do I say this? For people who are new to traffic generators, especially to layer 4 thru 7 traffic generators, I find there is sometimes confusion about tests. For example, I might have a customer ask me to run a “bandwidth test”. That is it…. just ‘can you run a bandwidth test’ with no additional details. I think looking at how the “two big” traffic generators companies have set up default starter tests would help maybe shed some light into this confusing area for people.

Since in this YouTube I was using Spirent, we will just take a sneak peek at their “starter test templates” while we are in there. Look at the varying modifications they have for the varying types of tests you might be asked to run. And I think a great place to start, for people trying to understand tests like CPS (connections per second), CC (concurrent connections), and BW (bandwidth), might be simply by looking at what defaults Spirent chose for some of those “start test templates”.

Below is just a quick table of 6 of the tests. Remember that “run a bandwidth test” I mentioned before? Well the top 3 starter templates ARE “bandwidth tests”. And they are tweaked differently. Also look at that MAX HTTP Connections per second test starter template CyberFlood has. HTTP 1.0 is being used with piggy-back get and also a RST instead of a FIN. What is a piggy-back get? What is a RST? :). Run a trial run of any of the below tests inside of the CyberFlood and you can download and look at the pcaps. :). Great great great way of learning!

Summary

So those are pretty much big ticket items I go into in the video. :). Enjoy! So so so much more to come!

YouTube

Categories: CyberFlood