# How We Helped EzTenda Cut Latency with GDN

When our client EzTenda expanded beyond the UK, performance issues quickly followed. Latency in Australia hit 7–9 seconds, threatening user experience and scalability. As their long-term tech partner, we stepped in with a solution: a Global Delivery Network powered by Stellate. Here's the breakdown.

- URL: https://sudolabs.com/insights/how-we-helped-eztenda-cut-latency-with-gdn
- Published: 2025-04-30
- Updated: 2025-05-23
- Author: Roman Šelmeci
- Categories: Tech
- Language: English

[EzTenda](https://sudolabs.com/case-studies/ez-tenda) is our long-standing client, on a mission to improve the way the drinks and hospitality industries do business through digital tendering. Their expansion from the UK into Spain and Australia introduced a new challenge: maintaining consistent performance across geographically dispersed user bases. Latency issues, particularly in Australia, were impacting user experience, with response times reaching up to 7–9 seconds. To address this, we implemented a **Global Delivery Network (GDN)** using Stellate, a solution that not only reduced latency but also positioned EzTenda for future scalability. In this article, we’ll discuss the challenges we faced, why we chose GDN, the implementation process, and the benefits and costs of this solution.

## Understanding the Problem: High Latency in Distant Regions

EzTenda’s platform is built on a serverless architecture hosted on Vercel, with data stored in a single NeonDB database located in Frankfurt, EU. While static content was cached by Vercel’s CDN, there was no caching layer for our GraphQL API. This meant that every query had to travel back to the origin server in Frankfurt, causing significant delays for users in regions like Australia.

### Key Challenges

- **High Latency in Australia:** Response times peaked at 7–9 seconds due to the physical distance between users and the origin server.
- **Scalability Concerns:** As EzTenda grew, the lack of a caching layer for dynamic content posed risks for handling increased traffic.
- **User Experience Impact:** Slow response times led to frustration among users, potentially hindering customer retention and growth.

Our goal was to reduce latency to under 1 second while ensuring scalability for future expansion into other regions like the US and additional EU countries.

## Why We Chose GDN

After evaluating multiple solutions, including database replication and alternative caching strategies, we decided to implement a GDN through Stellate. Here’s why:

- **_Global Performance Improvements_**

GDN deploys CDN servers globally, reducing latency not just in Australia but across all regions where EzTenda operates.

- **_Seamless Integration_**

Stellate’s GDN integrates directly with our existing GraphQL API without requiring major architectural changes or database migrations.

- **_Scalability and Future-Proofing_**

With Stellate’s infrastructure, we can handle projected growth and easily expand into new regions without manual configuration for each location.

- **_Cost Efficiency Over Time_**

While the upfront cost of $249/month is higher than some alternatives, it reduces engineering overhead and ensures long-term value through improved performance and reliability.

## How We Implemented GDN

![](https://a.storyblok.com/f/117609/1949x863/2c399c5b28/eztenda-gdn.png)

The implementation process was structured into six key steps:

1. **_Stellate Setup_**

   We created a Stellate account and configured it to work with EzTenda’s GraphQL API.
2. **_Integration with GraphQL API_**

   Stellate was integrated as a caching layer for dynamic content, allowing frequently accessed queries to be served from edge servers.
3. **_Data Invalidation Process_**

   To ensure data consistency across regions, we implemented an invalidation process that updates cached data whenever mutations occur.
4. **_Performance Testing_**

   We conducted extensive testing to measure latency improvements and ensure data integrity across all regions.
5. **_Phased Traffic Migration_**

   Traffic was gradually redirected to the GDN-enabled API to minimize risks during deployment.
6. **_Monitoring and Optimization_**

   Using Stellate’s analytics tools, we monitored performance metrics like latency and cache hit rates to fine-tune configurations.

## Benefits of GDN Implementation

![](https://a.storyblok.com/f/117609/5704x4384/27df773f71/cache-hit-rate.png)

The implementation brought measurable improvements across latency, scalability, and global delivery.

- **_Drastic Reduction in Latency_**

With CDN nodes deployed closer to users in Australia and Spain:

Latency dropped from 7–9 seconds to under 1 second in Australia. Users across all regions experienced faster load times.

- **_Improved User Experience_**

Faster response times led to higher customer satisfaction and engagement, directly impacting conversion rates positively.

- **_Scalability_**

The GDN infrastructure supports future growth without requiring significant changes as traffic increases or new regions are added.

- **_Simplified Architecture_**

By caching GraphQL queries at the edge, we reduced the load on our origin server in Frankfurt while avoiding complex database replication setups.

- **_Global Reach_**

The same infrastructure improvements benefited users worldwide, aligning with EzTenda’s long-term expansion goals.

## Costs of GDN Implementation

While the benefits were substantial, implementing a GDN came with associated costs:

- **_Monthly Subscription Fee_**

Stellate charges $249/month for its GDN service based on usage metrics like bandwidth consumption and server capacity.

- **_Engineering Effort_**

The initial setup required development resources to integrate Stellate and configure caching rules effectively.

- **_Data Transfer Costs_**

Data transfer between regions incurs additional charges depending on traffic volume (e.g., $0.042 per GiB for Australia).

- **_Ongoing Monitoring_**

Continuous monitoring and optimization require time and resources from our team to ensure peak performance.

## How GDN Reduces Latency

![](https://a.storyblok.com/f/117609/5708x4428/d8a7dd5165/cache-hit-rates.png)

To understand how GDN achieves such significant performance improvements:

- **Edge Caching:** Frequently accessed data is stored on CDN nodes closer to users.
- **Request Routing:** User requests are routed to the nearest edge server rather than traveling back to the origin server.
- **Data Replication:** Dynamic content is synchronized across CDN nodes using intelligent invalidation processes.
- **Shortened Data Paths:** By reducing the physical distance data travels, latency is minimized dramatically (e.g., from 9 seconds down to under 1 second).

## EzTenda Positioned for Global Growth

Implementing a Global Delivery Network through Stellate has transformed EzTenda’s performance globally by addressing latency issues effectively. The combination of edge caching, seamless integration with our GraphQL API, and scalability ensures that EzTenda can continue its expansion while delivering an exceptional user experience across all regions.

While there are costs associated with this solution, the long-term benefits, such as improved performance, scalability, and user satisfaction, far outweigh them. With GDN in place, EzTenda is now well-positioned to serve customers worldwide efficiently as it continues its growth journey.
