Blogs

In our increasingly wireless world, three letters frequently pop up when the talk turns to VoIP in a call center: QoS, or Quality of Service. When call center operators look at building, expanding, or virtualizing their business, QoS is a term that often gets tossed around like a minor detail or an implicit assumption during discussions with SIP providers or resellers. This Explain It! prepares you for these conversations, so you understand the basics of call center QoS and know why it is critical, so every customer contact is consistently clear.

QoS May Not Mean What You Think It Means

When you first hear the term quality of service, or QoS, you’re forgiven for thinking it refers to a specific condition. If you’re asked about the quality of service at a restaurant, you might respond that it was exceptional and attentive, or dreadfully poor. Or if you take your car in for an oil change, they might email you afterword with a survey about the quality of their service, where you reply they were cheerful and efficient, or that they filled your radiator with string cheese.

However, in the SIP space, QoS doesn’t refer to a condition. It is instead a prepared response to certain conditions. Think of it as a thermostat for your data network. The thermostat continually monitors the temperature of your home or office, so that when the air gets frosty, the heat kicks on. Or, when the summer sun blazes, it activates the air conditioner. You may also set it so that it doesn’t run while you sleep, or program it to heat up the house thirty minutes before your alarm clock goes off.

QoS is a protocol that monitors and prioritizes data traffic so that when things heat up in your network, users are guaranteed a certain level of service based on data flow. As the number of users and devices connected to your network expands, and the consumption of data increases, call center QoS plays a pivotal role in making sure that the streams of voice data are always given special treatment. Because, without call center QoS, conversations can quickly get mangled and garbled, which leads to a poor QoE (quality of experience).

QoS Keeps Them Packets Rolling

Many of us have experienced the dismay of receiving an email with a 15 GB attachment weighing it down like an anchor. If everyone in the company is copied on the email, the network groans under the strain, producing the worst kind of data constipation for those operations with limited bandwidth. It can also cripple call centers using VoIP since each call is competing for space to send their packets. If those packets are lost, slowed, or interrupted, you can consider that call toast.

With voice traffic, the sequence of those packets is non-negotiable. B always comes after A, and right before C.  As we’ve discussed in previous articles, conducting a VoIP conversation using SIP requires a high level of precision, since it involves two streams of data packets passing between each caller. (One stream sends media packets and the other receives them), and a third signaling stream to establish, maintain, and end the call. To top it off, the timing of those packets has to be nearly note-perfect, or it won’t sound like a real-time conversation.

To keep those packets rolling, administrators usually measure three key metrics:

• Latency: This captures the amount of time it takes for a packet to travel from Point A to Point B. Latency for most data traffic (anything from instant messages to emails) can vary without causing trouble. Latency for voice traffic, however, needs to stay below 100 milliseconds, with the ideal time being around 40 ms. If packets get slowed down trying to enter or leave a network, this can be a call killer.
• Dropped Packets: During times of heavy network congestion, packets can be dropped and lost, which slows down apps or message deliveries. However, those packets can be retransmitted, so everything eventually comes through. However, voice packets can’t be retransmitted, so callers are forced to endure poor and choppy calls.
• Jitter: Whereas latency is measuring travel time, jitter measures the delay between packet deliveries. A packet's delay varies with its position in the queues of the routers along the path between source and destination. As jitter increases, video conferences and VoIP calls degrade considerably

It’s an amazing technological feat when you consider the things that have to go right to coordinate the passage of voice packets across different lines, in and out of private and public networks, and through various routers and PBX systems to arrive in your agents’ ears in real-time. Call center QoS helps safeguard this minor miracle by prioritizing voice traffic in case an accidental cyber saboteur decides to torrent the entire library of The Simpsons or sends a .zip of their favorite Grateful Dead concert to everybody in their Contacts file.

Call Center QoS in a SaaS World

The examples of network killers we’ve mentioned are less frequent and less disastrous nowadays, due to increased capacity (which has replaced bandwidth as the metric which matters, since it measures the speed along the entire path between locations, not just your network). However, with the rise of SaaS applications, the total number of programs connecting to the cloud keeps rising, which can lead to a death by a thousand cuts scenario if your call center QoS is not accounting for them.

For an inbound call center, your agents may have numerous applications pinging their respective servers throughout their working day. You may have Salesforce as your CRM, Zendesk for customer support, 3CX for unified communications, Office365 for productivity, and Skillsoft for online training.  Call center operators depend on QoS to prioritize voice traffic, but it’s also essential for stack-ranking business-critical applications so that during periods of high congestion, 3CX and Zendesk aren’t bumped by a non-critical update from Microsoft or a training video from Skillsoft.

To properly manage call center QoS within these complex SaaS environments, the network administrator must configure the rules appropriately, which is the equivalent of programming the thermostat. This involves adjusting the Differentiated Service Code Point (DSCP) for each type of traffic, ranging from 0 (lowest priority) to 46 (highest priority). Since voice traffic is the top priority, it is set at 46, which ensures expedited forwarding. This is the equivalent of VIP status, placing this traffic at the front of the line. Meanwhile, a service like Spotify might be placed in the 0 category.

Tagging and prioritizing all of the types of data flowing through your call center takes time, but it provides two significant benefits that bolster QoS and QoE. The obvious payoff is in making sure that surges in usage on your network don’t produce jitter, latency, or packet loss. But the second benefit helps you from a planning and operational perspective since you’ll have a much clearer picture of the SaaS and cloud-based applications attached like tentacles to your network, and how much capacity they take up.

Bandwidth ≠ QoS

In previous articles, we’ve explained how you can calculate the amount of bandwidth you’ll need to support various VoIP powered call centers. This is a critical step since bandwidth caps the total number of simultaneous calls. But it doesn’t solve the problems of jitter and latency. Call center QoS is what guarantees that your agents will be able to communicate clearly to your clients, which improves the overall QoE.

The most significant issue harming the quality of voice traffic is poor design and configuration. Following best practices of call center QoS, you can prevent problems from occurring by prioritizing traffic, while monitoring uptime and usage patterns.

We take call center QoS seriously, so contact us today if you need a partner to help you architect the right call center solution for your business.

Connect with Voxtelesys on Facebook, Twitter, or LinkedIn.