Article by Thousand Eyes principal solutions architect Mike Hicks.
A lot of IT and digital teams learned to move at pace over the past few months. A largely unsustainable pace, it must be said, but one born out of necessity.
By most accounts, no business was spared and most compressed years of transformative impact into a few short months.
Now, the burning question is: to what extent did they get their setups right?
This is a live question when it comes to unified communications. Apps like Webex, Zoom and Teams became core to business operations overnight. Only now, with some breathing space, is there a chance to optimise how these video collaboration apps run in the enterprise environment to ensure business needs are met in the mid to longer-term.
When it comes to performance, platform vendors provide their set of definitions of what good should look like.
Microsoft, for example, identifies several metric thresholds they recommend to ensure the optimal experience for Teams end-users. These can be used as ‘baseline’ references to validate each Teams’ component’s performance.
Microsoft also has an approved provider scheme that recommends telecommunications carriers with optimal routing for Office 365 traffic.
But businesses can’t just rely on their platform or ISP to optimise access to unified communications services.
Given the now critical function of this technology, businesses themselves must also review and optimise their own setups for longer-term use.
Higher speed, higher performance?
There’s a long-held belief that higher speed internet services are needed to make web-based applications like unified communications perform at their best.
But, a business or user’s connection speed to the internet is not the critical determinant of an application’s or cloud service’s performance.
Performance is impacted by many factors, including where the unified communications servers are located concerning the business or end-user, and the number of hops – or devices – through which traffic passes on its round trip back to the user.
A recent study of this by the Australian Competition and Consumer Commission (ACCC) found “video conferencing applications hosted in Australia connected more quickly than those hosted overseas”, with as much as a tenfold difference in latency experienced by users the farther data packets had to travel. “Longer delays, or higher latency, are likely to lead to disjointed conversations during a video call,” it said.
In other words, users’ access to unified communications services was impacted more by distance from servers and less by the speed of their internet connection.
In addition, with bandwidth at a premium during pandemic-related lockdowns, some service providers temporarily scaled back the quality of services so they would function better on low-bandwidth or otherwise congested connections.
Though some of these restrictions have since been lifted, it shows these services can function on regular internet connections and without the highest video resolution possible.
Monitoring and testing
Businesses can do more to optimise the end-to-end unified communications chain, and get early warning of potential bottlenecks before they can disrupt the performance and usability of services.
The way employees connect to enterprise unified communications services has permanently changed.
Many will be accessing the services from home-based connections for the foreseeable future, and businesses and network administrators need to understand how to factor that extension of the network edge into our thinking, policies and architecture.
As we move into longer-term change, the gaps are more apparent.
What performance are we getting now? Are we getting what we expect to get? What can we now do to get better performance every time we use our business-critical unified communications service?
This kind of gap analysis or lifecycle approach is crucial to understanding how to optimise the network. Performance will not improve by looking at only the network or the application in isolation.
Moving away from siloed monitoring, and looking at performance end-to-end is crucial. Conducting a complete transaction analysis can help in this instance.
Businesses will also benefit from having some form of always-on, pre-fault testing. This can be used to identify a problem before it causes service degradation and impacts the productivity of end-users.
Recognising a potential problem as early as possible means having time to route around it and optimise performance.