Salesloft's Network Test analyzes your network's performance and provides clarity around the health of your network environment. A healthy network environment is vital to the performance of tools, such as the Dialer, as it allows them to operate at optimal efficiency.
We recommend completing a Network Test whenever you experience any Dialer-related quality issues as the test provides an insightful troubleshooting baseline.
Things to Note: When reviewing your Network Test results, any orange or red results indicate that your network connection may be unstable or suffering from other issues that could very well impact the Dialer performance.
In this article, we will walk through each aspect of Salesloft's Network Test and the significance of your results, including:
If you have any additional questions after completing a Network Test, we recommend submitting a support case for assistance.
Throughput can be defined as the available bandwidth your machine has access to on the directly connected network. This variable can fluctuate at any given moment so bandwidth speeds are typically measured in multiple captures and then averaged out to get the closest number.
Below is a quick snapshot of what a normal Throughput result would look like. Here we can see the minimum speed is roughly 36.5mbps after converting the 36,504kbps. On the maximum end, we can see a speed of 92mbps was reached which gives us an average of 67 mbps.
You may be asking, how do I know if these numbers are good? Well, this baseline will vary based on your actual provider bandwidth which we will revisit in a later section. If your internet speed should be 100mbps, but your machine is only measuring 1mbps, then that’s a good indicator your connection may be suffering.
As mentioned previously, taking the measurements of both the low (minimum) and high (maximum) is a reliable way to catch any performance issues. While the average will give the most accurate idea of your speeds, seeing a wide variance between the low and high will indicate an unstable connection. This can negatively impact the Dialer functionality and call quality as you may experience delays or poor audio if the network is unreliable and slow to respond.
Above is an example of an unreliable network measurement. Notice the red flags on the session count which tells us during the test there was almost no connection detected at all. Additionally, we can see the average speed was 0.158mbps which is almost a non-existent connection and will not handle a VoIP call reliably. It is very likely that the bandwidth speed will also register low and this particular network is severely underperforming which will in turn impact the Dialer experience.
Call quality is a rough measurement taken to assess how a call may sound at any point in time. It’s important to note that other variables go into call experience such as signal strength, connection, hardware, recipient, etc. that all differ on a per-call basis. Our network test simplifies this by breaking down a call into a few categories to give us an idea of how it may sound.
Below is an example of a good-sounding call followed by a breakdown of each measurement.
MOS Score registered a 4.4 which is considered “Excellent”. MOS or Mean Opinion Score is an industry-standard used to assess the quality of a call. The scale ranges from 1 - Bad, 2 - Poor, 3 - Fair, 4 - Good, & 5 - Excellent. Since reaching a Perfect 5 is typically impossible, any score between 4.3 - 4.5 would be considered an excellent range.
Round Trip is the time it takes for the call to terminate to the destination and return to your machine. The lower this value the better as that will mean faster delivery of voice traffic. In this case, 16ms is an exceptional time and typically anything <100ms would be considered acceptable.
Packet Loss is important as it indicates whether the voice test suffered from any missing audio data. Voice traffic is generally considered “best effort” and is not guaranteed to be delivered reliably. If you note any packet loss, that will likely indicate the audio may very well be choppy or delayed as the actual data is being lost in transit.
Jitter is the deviation from a normal time delay or signal. In this case, it will typically show up if high round trip times (pings) are recorded or packet loss is detected.
In our next example, we will look at what contributes to a poor-sounding call.
In the example above, we can see the Outbound MOS Score registered at 1.6 which is considered “Fair”. However, we can see the Outbound Packet Loss registered at 15.5% which will significantly impact the outgoing conversation. Due to the higher measurements taken, we can also see some Jitter present which tells us there was a slight deviation from what one would expect with an “Excellent” sounding call.
Other Call Quality Results
"Error: Connection Error (31005)"
In the above example, we can see no Call Quality measures were recorded in our result. This typically means our server was unable to communicate with the endpoint running the test. More often than not, this is the result of a blocked connection such as a firewall or cloud proxy limitation. While not pictured in this example, take note of the TURN Connectivity results as it’s very likely UDP, TCP, or TLS may be failing to establish which will further prove something on the network path could be preventing the traffic from flowing.
"Error: Could not establish call."
TURN connectivity is essentially a test for server responsiveness between your machine and our network. This test can be beneficial as it will test three different internet transport protocols and determine if any of them are blocked or have issues with the connection. This can be valuable for isolating any potential firewall or other network changes needed to help the Dialer connect for calling.
UDP is the most notable connection type to see as voice will almost always use the UDP protocol. What this means is that the voice communication between you and the recipient is considered “best effort” and will send the fastest rate of packets to deliver high-quality audio with minimal delay. If this test were to fail, it’s very likely there is a network configuration issue or adjustment needed to open up this communication.
TCP is similar to UDP, this is typically the backup transport protocol when UDP is unavailable. This protocol is slower but more reliable as it ensures every packet is received in the communication channel. While it may be more reliable, it is not ideal for voice transmission as real-time delivery relies on speed to minimize any conversation delays. If this test were to fail or be unavailable, it would also be indicative of a network configuration issue and should be reviewed along with the other protocols.
TLS is another transport protocol that provides a security layer for the normal transmission of data. The results here are not as significant as the UDP/TCP response times but act as another metric to average out the time it takes for your machine to connect to our servers.
Below is an example of what a healthy TURN test looks like:
Here we can see that all three of our transport protocols passed successfully. The measured time is the delay from your machine to the actual server and not the total round trip time. While the speed can be indicative of an unstable or congested network, seeing a successful connection will be more significant.
Above is an example of a failed network transport test in which the UDP protocol failed to connect.
Additionally, the TCP & TLS connection rate was measured extremely high. This tells us that there is either a blocked connection for the UDP protocol or the network is extremely unstable and failed to establish a connection with our servers for Dialer communication. In this situation, it would be best to review the network configuration to ensure no firewalls or security endpoints may be interfering with the connection along with any congestion resulting in an unstable connection.
Location represents exactly what it says, it’s our best guess of where you are physically located. Note, this may be inaccurate as there are other factors in play that may change your location such as VPN clients and cloud proxies.
Below is a quick example of a location detection from a network test. Here we can see the public IP address used for this test registered in Phoenixville, Pennsylvania and the service provider was Verizon Fios. This result is typically more of an informational data point and can help distinguish if any machines are registering from inaccurate locations or may be routing through other network/security endpoints adding onto any negative feedback.
Bandwidth speed is a metric many of us are familiar with as its the speeds internet providers agree to provide with their service. However, it’s important to note that the true speed measured at any given time may vary on other external factors. Verifying this metric on the network test is a useful way to determine if those factors may be limiting your available connection speed.
Below is an example of a typical network test that indicates the speeds do in fact match what was contracted. In this case, it is a Verizon Fios (fiber optic) connection that should be reaching around 250mbps up & down. We can easily determine that is the case and there is minimal bandwidth saturation which is great to see. You will also note Jitter is being measured which is present to help indicate if there are any obvious indicators the network is congested. Typically we can expect to see jitter register higher as the speeds measure lower.
Below is an example of a bandwidth measurement where the connection may not have been completely available or significantly impacted.
We can see the upload speed registered at 0.6mbps and the download speed registered at 36mbps. As a consequence, Jitter (400ms) was extremely high which further indicates this connection is suffering from latency delays and other connectivity issues.
An important thing to note with bandwidth measurement is voice transmission requires a reliable upload and download. Through our typical use of the internet, we mainly visit websites and download information to save or display. However, since voice is a two-way communication, it also requires a solid upload to ensure our data is reaching the recipient. If you notice your bandwidth speed is suffering on either the upload/download, then it could very well indicate congestion or an unstable connection that will impact the two-way audio transmission.
Ping is a common network tool intended as a basic measurement of the time it takes for one network endpoint to respond to another endpoint. This can span anywhere from within the same network to around the entire globe.
With our network test, we take a quick measurement of the time it takes for your machine to ping the network server it would connect to for Dialer services. We then record the time it takes for a response to come back to your machine which can be a useful and quick way to determine if there are any delays or issues with that connection.
Below is an example of a Ping test that was located in the US East region. We can see the response time was only 14ms which is a great time as we want to see response times <100ms domestically.
Below is an example of a Ping test that had a very poor response time resulting in heavy delays with communicating to the server. In this case the region tested was Sydney and the response time was 406ms which would be considered extremely slow in networking.
Machine information will display device related information that could impact the Dialer performance. This is a data point used more for support related troubleshooting to determine machine type and specs. Below is an example of the various data points we gather from a user’s machine. We are able to view the CPU capacity and utilization which can impact performance when under load. Additionally, we can determine the manufacturer and model along with the operating system version. This can be helpful in isolating patterns or widespread issues with specific builds or updates.
Note, the above information is only collected if the optional testRTC Companion extension is installed on the machine running the network test. Having this extension installed is not a requirement to complete the Network Test and if it is missing from the browser, this widget will display a message indicating so. A link to the extension will be referenced within the widget and once it is installed, the test may be executed again to gather Machine Information.
Device state, similar to machine information, is another beneficial way to gather information around the type of environment the Dialer is operating in. Additional data points here include microphone access, wireless or wired connection, and the type of audio device being used. These are all helpful troubleshooting data points to better understand the type of network connection & hardware being used for the Dialer in case specific variables could be contributing to the type of issues experienced.
DNS Lookup is another way for us to measure the time it takes to reach various servers globally. We have a list of 9 different network points to communicate with, some related to Salesloft services and others publicly accessible. This helps us identify any connectivity issues or delays within the platform itself and with the regular internet. Below is an example of what a normal response time would look like.
Below, we are able to see that some of the results took a significant amount of time to respond to. While this would be considered lower priority, it does help highlight any abnormal network conditions that could be contributing to network congestion or reachability issues.
Network Score is the cumulative grade of all network components and their results. This will average out the performance of each widget and then provide a summary list of conclusions and recommendations regarding the current network connection. Below, you can see an example of a network test that performed well and had a good score. While the score was considered “Good” there is always room for improvement to make things optimal as this is only a snapshot of an ongoing connection that can fluctuate and change at any moment.
Below is an example of a very poor Network Test and Score. All of the components returned a very poor result and therefore, the network itself is in “Critical” condition. Again, we highlight all of the areas that suffered and ways to improve them. The areas that require the most attention are highlighted in red and usually correspond to any specific widgets that resulted in a red threshold or error. These would be the areas to prioritize as far as improvements go but it is still equally important to review all suggestions as making adjustments in various areas can contribute to an overall performance improvement.
Now that we have gained some additional insight with our network test tool, the biggest question remaining is where do we go from here?
Below are a few recommended next steps to assist with proactive troubleshooting outside of our Support channels.
Things to Note: These steps may not be necessary or applicable to every network environment, but could still prove to be helpful in finding a speedy resolution.
- Disconnecting/reconnecting to the existing network, especially on a wireless connection in case signal strength is weak.
- Switching to a wired connection if using wireless to see if any reliability improvements are experienced.
- Connecting to a different network if available, typically public networks or even a mobile hotspot can be useful for isolation tests.
- Restarting the machine & router to see if any performance issues clear up, sometimes devices can bottleneck unseen processing issues and a fresh reboot will purge the system out.
- Setup Passthrough if possible to determine if the call quality issues may be on the recipient end which can also be a helpful way to rule out the Dialer and network.
- Proactively complete many network tests at different times of the day, on good days, disconnected from VPNs, or on different networks and hardware. This is a great way to establish a baseline and find patterns in the different components/factors being tested.
If the above steps do not improve the Dialer experience then we always recommend reaching out to our Support team to get a ticket created for the troubleshooting process. Our agents will assist with next steps to help identify any potential issues based on network test results and other data points.
