Internal Corporate Communications have never been more prevalent than at the current moment. CEOs nowadays expect to have a direct line to their entire workforce at the push of a button, making “all hands” events far more frequent than ever before.
From a deployment standpoint, you want to control and test all the different variables that could go wrong at any point of one of these massive streaming events: for example, the playback might suffer from CDN failure, an encoding problem, a misconfigured event page, authorization issues, or what’s even worse, network overflow resulting in the throttling of the entire network. A throttled network could effectively stall productivity and basic services on a company-wide basis.
The only way to mitigate most or indeed all of these potential errors is to organize dry runs of any event before it takes place. By simulating an event and stress testing the network beforehand, we are able to identify the true capacity of the network, the expected video quality, and single out problematic portions of the network or even individual devices that require intervention.
But it’s not as simple as that.
On a network with hundreds or thousands of connected devices, obtaining infosec approval and subsequently deploying and configuring tester software is a logistical nightmare in and of itself. Additionally, running a video rehearsal on all of the company’s devices can not only be intrusive but can end up impeding productivity. The easiest way to mitigate the disruptiveness of these tests is to try to stage the tests during off-hours, but this ultimately undermines the underlying conditions of the test (network conditions at 12AM are not the same as network conditions at 12PM), thereby diminishing the test’s effectiveness.
With this in mind, Peer5’s engineering team has developed a super-simple silent tester that requires no software installation by the network administrator, no action on the part of the user, and actually runs transparently in the background on as many or as few target devices as desired. This makes it easy to perform as many tests as necessary before an important event, under varying conditions, or even to only troubleshoot a subset of devices, whenever needed.
What is a silent test?
A silent test is a simulation of a video broadcast that runs in the background of the actual devices on your network. The users are completely unaware that the test is occurring. It’s essentially a test run of the network and client devices that allows stress-testing and troubleshooting before an actual streaming event.
The only difference between a silent test and the real thing is user participation. With the silent test, the user is not at all inconvenienced, which allows for more extensive and effective testing. You can test and test as often as needed until full optimization is achieved.
How to run a Peer5 silent test
To set up a silent test, all you have to do is access the app.peer5.com Admin Console, select Silent Testing in the left-hand sidebar and then set the desired parameters in the easy-to-use interface as shown. Let’s walk you through the process:
The first step is to make sure that all devices meant to run the test are both turned on and connected to the network. The devices that execute the silent test are called runners, and can be integrated via a Chrome Extension, which can be deployed centrally using Chrome Enterprise or Windows Group Policy, via SharePoint or PowerShell integration, or through other custom means.
By “runner” all we mean is any device on the network that is participating in the test. Their participation is passive, in that, as mentioned above, all they have to do is be turned on and connected to the network. The test will run in the background, silently. The only way the user of the device will be aware of the test is if the stress test actually maxes out the network, causing connectivity issues.
Peer5’s engineers have also come up with an alternative called a “direct runner” which, unlike the Chrome, SharePoint and PowerShell options mentioned above, only requires that the tester leave open a Chrome tab with a particular URL. This URL is device-agnostic, so it can be opened ad hoc on any type of device that is connected to the network and has a web browser. Instead of running 100% invisibly in the background, it utilizes an open browser tab (the test is still silent, as the video will not be visible). The only requirement here is that Chrome must remain open and the device must remain connected to the network. To access the universal direct runner URL, select the “?” on the top right of the Silent Testing page.
Once connected, all runners show up under the Connected Devices list on the right of the Silent Testing panel. They will remain idle until a silent test is scheduled to run.
The second step is to schedule a test by clicking on the “+” button. All you have to do is set up a few basic parameters, starting with a timeframe (20-30 minutes minimum recommended).
By selecting “Built-in stream,” the test will simulate the broadcast of a generic video stream provided by Peer5. “Custom stream” will play a video defined by the customer, and “Custom page” will stream the customer’s actual landing page. Please note that this last option requires autoplay to be activated, as there will be no user to hit the play button. The page must also have Peer5 integration with the correct token. We recommend setting the audio to mute by default, as this setting removes certain Chrome limitations on autoplay videos.
Finally, we select the desired video quality, protocol and start delay. The start delay setting distributes the onboarding of test users over a given period of time. So if you’re testing 1000 devices, they will be connected randomly over the time interval selected (over 60 seconds, 5 minutes, etc.) in order to simulate real-world circumstances and avoid artificially crashing the system with 1000 simultaneous connections, for example.
When the status goes from Pending to Ongoing, runners start executing the simulation.
Once a silent test has been initiated, you can break down the results using the same industry-leading analytics tools Peer5 provides for actual streaming events, both real-time as well as historic. Our next post will provide you with a walkthrough of our Analytics tools.
Silent Testing Recommendations
Make sure the test doesn’t coincide with other video streams.
For high scale stress tests, if you're not sure about the capacity of your network, start testing on the weekend or during idle times (just in case you overload your network). You can also do such tests gradually, where at first you target smaller sets of devices and gradually increase the test sets.
Start with lower bitrates and gradually increase.
At least 100 users are necessary for any test results to be representative of a real event. For the test run try to include as many users as possible, simulating circumstances as close as possible to the real thing.
Minimum broadcast time should be at least 20-60 minutes, as if the test is too short it won’t properly simulate the real-world conditions of viewers connecting incrementally or the peering capabilities of the platform, as peering also requires a ramp up.
The dynamic and discreet nature of the silent test allows you to do as many tests as desired. You no longer have to rely on one large, disruptive test. After each test you can review and refine and test again. Peer5’s product is plug and play, as easy to implement as conceivably possible, but even so networks are complicated things. Our silent tester, hand in hand with our analytics tools, is designed to help identify the source of any problems before they become apparent to any users on the network.
To request a demo, please click here.
To take a look at our dedicated Silent Tester page, click here.