CHWC deployed twenty six Sapphire Eye sensors throughout their facilities. Once installed, the Sapphire Eye sensors connected to the APs around them, and continuously ran a series of passive and active tests. The active tests emulated client experience – attaching to each AP, authenticating, receiving an IP address and running background tests to measure performance metrics such as upload and download throughput, delay, packet loss, jitter and voice quality (MOS). This quantified the network’s ability to deliver the Wi-Fi service. The passive tests sampled all the other client to AP interactions, tracking key metrics such as connection data rate and retransmissions, thus quantifying each client’s ability to use the Wi-Fi service. Sapphire Eye test results are then analyzed by the Sapphire Analytics Engine and presented in a browser dashboard as Key Performance Indicators (KPIs) which make it easy to identify the areas of potential concern. CHWC also asked 7signal to perform an Optimization project to baseline the initial performance, suggest improvements and then track the effect of changes.
"We're not running blind anymore. Now, we can tackle performance issues proactively, before users feel it, not after the fact."
After a week of data collection, the Sapphire system had a good baseline of the CHWC performance. This could be shown in the Service Level Agreement (SLA) View where red, yellow and green is used to indicate compliance against target values for KPIs. The baseline SLA View (see Figure 1) showed there were issues with attach rates and throughput in both the 2.4GHz and 5GHz bands.
This provided the level of visibility into the end-user experience that CHWC wanted to see. Using the Sapphire software, CHWC and 7signal worked together to identify a series of steps to improve performance, including adjusting the channel plan and AP power levels. After each step, the team could see the impact of the changes and assess the performance improvement. The SLA view after optimization (Fig 2) showed significant improvement in all metrics.
Another helpful way to examine performance is to look at a key performance indicator such as throughput over time. Not only can one see the impact of performance on certain days of the week or hours of the day, but also one can view the impact of the changes. Figure 3 shows uplink throughput for the Archbold Medical Center -- the result after several configuration adjustments was a 60% improvement in some areas. In parts of Bryan Hospital, the improvement in worst case performance was even more dramatic at 200%!
The system also provided visibility of the Wi-Fi experience for individual clients. Figure 4 shows how the retransmission rates of individual clients were reduced dramatically over the course of the optimization project. Declining retransmissions are indicative of more airtime available to transmit and receive more data in shorter times, which makes a positive impact on the user experience.
The CHWC team now has the visibility it needs to proactively manage its wireless environment. In addition to the intuitive dashboards and graphs available from Sapphire, the system has been set up to generate automatic performance reports in pdf form that are automatically emailed to the IT management team. The team has also enabled alarms which generate automatic alert whenever certain KPIs fall below an acceptable threshold. In this way, the IT team is confident they will always be ahead of any issues that may arise as the environment changes, new applications are introduced or new devices join the network.