A recent report by STL Partners emphasized the significance of one of the three important subscriber criteria in judging a good first connection impression – latency.
Latency is the round trip time it takes a subscriber to see/hear the response to an action initiated on his device. Latency determines the subscriber experience of service provider network performance much more than download speed (throughput).
Latency is essentially the response time after a subscriber initiates an action on his device –
- the amount of time it takes for the page to load, or
- the wait before you hear ringing for a voice call, or
- the wait for the streaming video to load and play, etc.
Latency can be caused by many things besides the network itself – physical distance from key servers, misconfigured apps, local emergency situations that cause congestion, unresponsive application servers or underperforming DNS (within the operator network or externally). All of these will cause the subscriber to perceive the network as “slow” and increase levels of subscriber dissatisfaction and potential churn, even though many factors are outside the operator control.
Service providers however strive to minimize latency wherever they can. Network operators rigorously invest to drive down excessive latency in their networks and continuously measure latency within their network. To meet escalating traffic demands and in recognizing the significance of latency on the user experience, the mobile industry standards association 3GPP developed LTE-Advanced (releases 10-13), providing major enhancements of the LTE standard deployed in many mobile networks today. This technology targets peak data rates up to 1 Gbps and introduces new RAN concepts with the ultimate goal of designing a system that is drastically enhanced in both cell capacity and coverage. In LTE Advanced, end-to-end latency targets are substantially reduced down to 10 ms or less. This is especially critical to support the high growth, latency sensitive mobile applications for machine-to-machine / sensory data and interactive mobile gaming.
STL ranked service providers across Europe by average latency, using data collected from millions of Crittercism mobile applications – anonymized and aggregated. STL measured the total round trip time from when the subscriber touched the button on their device to when they got a response. They defined roundtrip latency under half a second (<500 ms) as “satisfactory” – anything over that was not. Latency under 100 ms is perceived as being “instant”.
An especially interesting point was the discussion about how irrelevant “average latency” is to the subscriber experience. This means that even if the average latency is within an “acceptable” range or even higher than target, if 20% of the individual subscriber requests are unacceptable (500 ms+), that subscriber will be more likely to be dissatisfied. To put it in the context described in my earlier blog: if 30 out of the 150 daily “requests” from an individual subscriber smartphone experienced latency of 500 ms or more, the subscriber will likely be dissatisfied – this is true even if the other 120 requests were faster than average.
This speaks highly of the importance of having robust, carrier grade technology to ensure that one of the first connection steps – the DNS request – is consistently ultra-fast, even when under any number of threats or traffic conditions. Infoblox pounds out latency in its solution design to eliminate unneeded recursion and to prevent DDoS attacks from impeding performance of legitimate requests. The Infoblox solutions for service providers provide consistent low latency, fast response.
At Infoblox, we strive to help service providers make a better first connection impression – every time. Oh, and by the way, of those service providers listed by STL Partners with the best latency performance… many are Infoblox customers.