Shuffll

Improving Shuffll Performance in 80%

Improving Shuffll Performance in 80%
Industry:
SaaS
Location:
Israel
Features:
Performance Monitoring, Traffic Auto-pilot
79
%
P-50 TTLB Improvement Worldwide
66
%
P-50 TTLB Improvement in US

Overview

Shuffll enables its clients to scale their video content strategy. Shuffll’s platform allows for the creation of end-to-end video content quickly.
Shuffll utilizes IO River to virtualize the CDN layer and execute a Multi-CDN strategy that includes Azure and Fastly.

Challenge

The primary challenge for Shuffle is to achieve performance optimization at scale.
The transition to a Multi-CDN architecture requires the capability to gain visibility over all traffic in order to start optimizing it. Shuffll difficulty is the ability to develop visibility over entire traffic, measure performance, and compare different CDN providers, with the option to drill down into the sub-metrics.
Such capability requires extensive and complex development, which IO River directly provides off the shelf.

{{story-testimonial}}

Shuffll's DevOps team is highly sensitive to TTLB (Time to Last Byte) due to Shuffll’s need for fast downloading of static files. The team is making continuous efforts to reduce the TTLB.

Improving Shuffll Performance in 80%

Solution

Shuffll leverages IO River's Performance Monitoring tools, which allow for performance comparison between different CDNs in different regions and at different times. This system identified a significant 80% performance gap between Azure and Fastly.

Using this system, the Shuffll DevOps team discovered that the use of a short TTL resulted in the cache being cleared every 3 minutes, with Azure facing greater challenges in handling this compared to Fastly.

Shuffll’s DevOps team made the decision to empower the Performance Monitoring system to autonomously handle the situation. By implementing a straightforward Dynamic Traffic Management Rule, IO River’s platform was able to determine the optimal CDN for each user at any given moment.

Results

  • P-50 TTLB Improves by 79% worldwide
  • P-50 TTLB Improves by 66% US

"Finally, we can see the whole picture, identify performance gaps, and easily optimize. It was like driving a Ferrari in the city. We gained visibility that allowed us to find the paths that enable boost performance for our users."

Issac Gil
CEO and CTO @Shuffll
Shuffll