In the dynamic world of fast fashion, delivering the latest trends to customers worldwide requires a seamless online experience. PrettyLittleThing (PLT), a leading fast-fashion brand, faced the challenge of optimizing their website's caching architecture to enhance speed, scale, and cache efficiency. To overcome these hurdles, PLT turned to Varnish Enterprise, a premium caching and content delivery solution. The move resulted in remarkable success, showcasing a significant improvement in server efficiency and load handling.

The Challenge: Speed, Scale, and Cache Efficiency

For an e-commerce giant like PLT, the efficiency of content delivery and ecommerce services is paramount. Caching, particularly during heavy load situations, significantly impacts site performance and page response times. PLT initially leveraged the open-source Varnish Cache to achieve faster content delivery and reduce backend server workload.

As PLT's traffic load increased and their technology stack expanded, they encountered slower responses from Varnish Cache servers, along with memory and performance issues related to bans and the ban lurker process. To address these challenges, PLT sought a solution that would enable them to scale efficiently while maintaining optimal performance.

The Solution: Migration to Varnish Enterprise

PLT transitioned from Varnish Cache to Varnish Enterprise, a move driven by the need to scale and adopt additional features. Adeel Ejaz, PLT's Group Technology Director – Digital Innovation, highlighted the importance of Varnish Enterprise's performance-oriented approach: 

"Performance is at the heart of what Varnish does, so we knew we needed support from Varnish engineers directly to help us debug our performance issues and assist with building a highly scalable caching solution."

Varnish Enterprise provided PLT with the flexibility and support needed to overcome performance challenges. Key features such as Varnish Broadcaster and Xkey were instrumental in enhancing cache handling. PLT successfully replaced Varnish bans with the Xkey feature, allowing for a custom implementation that better aligned with their specific requirements. Moreover, they utilized Varnish Broadcaster to streamline the update process for Varnish servers with purge requests, incorporating them into their autoscaling solution.

Results: Improved Performance and Efficient Scaling

The migration to Varnish Enterprise yielded impressive results for PrettyLittleThing:

1. Reduced backend load and improved performance. PLT experienced a significant reduction in backend load, resulting in improved overall performance
2. Servers handled 5x more traffic. Varnish Enterprise empowered PLT's servers to handle a fivefold increase in traffic, for enhanced scalability and efficiency.
3. Better hardware and memory utilization. By reducing the load on the ban lurker process, PLT achieved better utilization of hardware and memory resources.
4. Enhanced uptime and stability. Server performance and bottleneck issues were successfully resolved, leading to better uptime and a more stable service.

In conclusion, Varnish Enterprise not only met but exceeded PrettyLittleThing's expectations, enabling them to navigate varied traffic loads seamlessly. For technology leaders, DevOps teams, and site reliability engineers seeking a caching solution that combines speed, scalability, and efficiency, the success of PrettyLittleThing with Varnish Enterprise serves as a compelling testament to the platform's capabilities. See the full case study here.

Embrace best-in-class caching with Varnish Enterprise and elevate your online performance effortlessly. Our experts are here to help: Talk to an Expert.