The anticipated spike in traffic was estimated to be relatively short in duration (approximately 48 hours with a very large spike in the first 3 hours of the event). Therefore, an elastic, scalable and cost-effective approach was required to ensure an immediate escalation of the infrastructure. The ability to quickly scale-down once the campaign had reached a conclusion was also a priority to maintain the integrity of the operations budget. In previous years, the Splash of Pink team had to rely on third-party services to throttle visitor traffic to the shopping site, resulting in multi-hour wait times for shoppers to complete their purchase.

The Splash of Pink Team needed additional guidance on how to fully assess what the heightened traffic pattern would look like, analyze the environment’s current capabilities, forecast the additional capacity needed and implement the infrastructure required to sustain the heightened influx of web visitors.

The site had to be assessed, and an in-depth analysis was performed to determine its ability to withstand significant traffic spikes. Armed with the data from previous sale events, JetRails engineers performed several stress tests which indicated that the site would not withstand traffic demands of the magnitude anticipated.

The current website was hosted on a dual server platform. A clustered solution was engineered with dedicated systems for the SQL database and the addition of seven load balanced web servers. A content delivery network was implemented allowing for immediate access to static content, optimizing web traffic navigation and speed.

The newly engineered platform was stringently tested, and the results confirmed that it would easily sustain 19X its normal hourly visitors. The platform was tested and put into production before the traffic spike. JetRails engineers monitored the platform during the event to ensure that visitors received a great shopping experience without any interruptions or delays.