Adaptive Codecs – Good for Users, Bad for Networks

by Sorell Slaymaker

The Quality of Experience (QoE) for cloud based, team collaboration tools is poor for many enterprise users. This needs to change!

Network managers have the responsibility of a providing a high performance and reliable network for the enterprise, but have less and less control of what runs on their network. Talk with any network manager about their top 3 problems and collaboration tools are on this list. As more teams choose freemium, cloud based collaboration tools, which utilize adaptive codecs for their voice and video sessions; this is exacerbating the “wave” effect on enterprise networks.

The wave effect on networks is when the network utilization hits 100%, backs off, then hits 100%, backs off, and this cycle continues. Every time the network hits 100%, all applications are impacted, and this impacts critical applications that run the enterprise. TCP windowing starts the wave effect with a TCP session creating a large window and then backing off once a few packets are dropped. Having many flows across a single network link leads to large waves. One thing that WAN optimizers due is to rate limit the TCP flows to avoid the wave effect along with prioritizing flows.

Adaptive codecs do the same thing. When bandwidth is constrained, they will reduce the amount of packets they are sending until the constraint is lifted, then they will go back to taking the amount of bandwidth they need, about 100Kbps for voice and anywhere from 256Kbps to 30Mbps for video. Adaptive codecs work very well when there are a few sessions running on a network. Get lots of sessions together on a single network pipe, and waves occur if the network pipe is not big enough. WAN optimizers do not solve this problem because they are designed for TCP optimization, not real-time UDP optimization.

Traditional network gear from Cisco and others do not solve the adaptive codec wave problem. While Call Admission Control (CAC) can limit the number of concurrent voice and video sessions on an internal collaboration system, it does not work on a cloud based solution. Furthermore, networks are designed to both control what goes across it and enforce quality of service through packet marking. Neither of these solves the adaptive codec wave problem.

Compounding the problem, some of the team collaboration tools are encrypted using TLS, other team collaboration tools use WebRTC which does the encryption natively. Both make identification of the voice and video traffic very difficult.

Some of the SD-WAN players such as Cloudgenix are building in the flow intelligence to be able to solve this problem in real-time. They can:

  1. Send traffic across multiple paths based on network packet loss, latency, and jitter
  2. Prioritize voice above video and rate limit the video as required, esp across networks that are not honoring the diffserv QoS settings
  3. Identify and control real-time traffic, even if it is coming from a cloud provider, based on the flow characteristics such as voice which has packets of a fixed size that come at a fixed increment of time.

One enterprise told me last week that Cisco’s Spark, Cisco’s team collaboration tool, is loved by its users, but their Cisco WAN cannot manage the traffic effectively, causing the voice traffic on a conference call to be intermittent. Users who work at home have a better experience than those in the office. Cisco’s I-WAN product, their SD-WAN solution, is not optimized to support WebRTC and prioritize voice above video and data within an encrypted WebRTC session.

Enterprises have 3 options to solve the adaptive codec wave problem from cloud based collaboration tools:

  1. Block it – Put strict firewall rules in place and try and block all cloud based collaboration tools
  2. Overbuild – Add big networking pipes so network congestion does not occur
  3. SD-WAN – Dynamically manage the voice and video traffic riding the network

Option 3 is emerging as the most popular since blocking traffic only makes IT more unpopular and overbuilding leads to spending too much. The emerging SD-WAN vendors can solve this problem elegantly with new ways of doing bandwidth shaping across multiple paths, while the mature networking vendors are pushing the overbuild strategy.