How does one ensure the smooth rollout of new, revolutionary, disruptive technologies such as VoLTE, especially when subscribers have high expectations for performance and reliability that evolved out of their experience with mature, legacy technologies?
Start with an old cliché: You can only manage what you can measure. However, most old clichés are grounded in truth and comprehensive, intensive, end-to-end measurement is the foundation for successful technology deployments. Reliable and repeatable performance test capabilities ensure that issues are caught and addressed before new technology is rolled out to end customers.
But for many reasons, this is not easy. As one example, during a VoLTE rollout, there are multiple new interactions which must be managed, including interactions between networks, devices, and the environment. Resolving problems to the appropriate domain and identifying root causes to the device, network, or interaction consumes considerable amounts of time and resources. By definition, innovative technologies are in their infancy and one should therefore expect that new issues appear frequently. Debugging can be quite complicated, often requiring systems engineering support which may create additional bottlenecks during rollout.
A new perspective on test is required. Two case studies highlight the value of device-network analytics. In both cases, the analysis focused on the device-network interaction which provided direct insight into the root cause of the issue.
Case Study 1: Video stalling issue
- Test need: Pre-launch testing a new smartphone
- Test application: Streaming YouTube
- Symptom: Device under test (DUT) displayed inadequate performance when compared to the reference device.
- Results: An analysis of device and network logs showed that not only was the DUT's YouTube performance worse, it was also signaling the network more. So, in essence, the device was delivering poor user performance and loading the network, i.e. the worst scenario for an operator. The TCP/IP settings on the DUT weren't optimal (for YouTube) thereby causing a TCP reset and a network reconnection.
Case Study 2: Chatty device
- Test need: Pre-launch test of a new smartphone
- Test requirements: Networks in which the device was being tested reported increased load on the network or network congestion though there were no signs at a device level
- Symptoms: The field team was not able to see an issue since the throughput of the DUT was comparable to the throughput of the reference device
- Results: Analysis of device and network logs showed immediately that while the two devices had similar (user) throughput, the DUT consumed twice as much network resource as the reference device to achieve the same throughput, thereby putting greater load on the network.
Clearly, isolated testing of either devices or networks is no longer sufficient to identify and troubleshoot performance issues. This problem will only grow as networks continue to become more complex. Rollouts of advanced features such as VoLTE will only go smoothly if the interaction of the device and networks while running the application is analyzed. Successful rollouts will require effective tools and integrated analytics for device network interaction will be essential.
— Vivek Vadakkuppattu, Director & Head, Analytics Business Unit, Azimuth Systems Inc.