Package.AI Optimizes AWS Infrastructure by Automating Reserved Instance Management

Key Challenges

Frequent capacity changes made it impossible for Package.ai to forecast commitments, resulting in high compute costs by running On-Demand

Key Results

Package.ai achieved a 95% increase in RI coverage.

As a result, they reduced EC2 costs by 42%.

The Challenge:
High EC2 costs

With a rapidly growing customer base and the need to accelerate product development and client adoption, Package.AI turned to AWS for an elastic cloud compute infrastructure that enables quick and easy scaling to meet the continuous rise in demand. 

As a startup company with a cost-efficient mindset in place, Package.AI began their AWS journey by using EC2 Reserved Instances (aka RI’s), in order to reduce their compute costs. However, due to the fact that the demand for their service is irregular, scaling up or down with RI’s quickly became a challenge. The engineering team of Package.AI came to realize that spending resources on constantly managing an RI fleet manually, in parallel to their mission to deliver a solid user experience and new product features, was simply not feasible. 

The main challenge Package.AI faced while managing their commitments, was the fact that they needed to react in Real-Time to their capacity changes, and doing that manually without reliable predictions rapidly became a technical and operational nightmare. 

We were looking for a no-touch solution that would manage our commitments in the most efficient way possible

Yoav Sadeh, CTO and Co-founder 

The Solution:
Automated Commitment Management

In order to free up engineering resources from cost optimization towards product development, Package.AI began researching a cost optimization platform, which will attend to their ever-growing operational requirements.   

During their research, Package.AI came across various cost optimization solutions, which in turn were very insightful towards their cloud spendings, but required them to undertake continuous manual actions, which they were attempting to eliminate

Managing commitments manually slowly became an operational burden, and overall wasn’t as effective as Zesty’s engine

Yoav Sadeh, CTO and Co-founder 

Eventually, Package.AI decided to partner with Zesty in order to tackle their automation challenges. Zesty has provided Package.AI with an automated cost optimization platform, which leverages AI models to analyze its entire EC2 deployment in real-time and react accordingly. 

According to real-time capacity changes, Zesty’s AI engine automatically buys and sells RI’s via the RI marketplace. This proactive and automated approach to RI management empowered Package.AI to implement Zesty’s solution shortly after the initial demo. 

Due to its non-intrusive architecture, Zesty was able to on-board Package.AI in less than 15 minutes, reducing the time to value to a minimum.

Zesty is a full-suite solution that covers everything related to cost optimization of EC2 AWS infrastructure with a user-friendly dashboard and a great customer-oriented service

Yoav Sadeh, CTO and Co-founder 

The Result:
Maximized savings, minimized waste

Besides freeing their valuable engineering team from manually dealing with cost optimization, Package.AI was able to achieve dramatic results and ROI from using the Zesty platform:

95% increase in RI coverage: Prior to implementing Zesty, Package.AI was able to cover only a small fraction of its infrastructure with commitments. Within days after the implementation, Package.AI witnessed a steady increase in the percentage of their infrastructure covered with savings and eventually reached the holy grail of over 95% infrastructure covered with savings.   

42% more savings with zero engineering efforts: Using Zesty’s AI engine, Package.AI was able to leverage the most lucrative available marketplace discounts, while eliminating any manual efforts which were required to manage commitments.

100% utilization while supporting all scaling scenarios: Due to its fast automatic operations, Zesty was able to support any scaling scenarios with perfect utilization, eliminating the pain of over-provisioning and the operational risk of under-provisioning

After realizing Zesty’s unique success-based business model, accompanied by a buy-back guarantee, it was a no-brainer for us

Yoav Sadeh, CTO and Co-Founder