Freightos.com, a Freightos Group Company, aims to be at the center of that change, by making global trade frictionless as the world’s online marketplace for the trillion-dollar international shipping industry. Freightos.com is a freight marketplace helping importers and exporters reduce logistics spend and save time with instant comparison, booking, and management of air, ocean, and land shipments from top logistics providers. The result is international shipping that’s as smooth as booking a flight online.
Founded in 2012, the company has six offices across the United States, Europe, Asia, and the Middle East and works with major retailers such as Marks and Spencer.
“Our marketplace offers an unrivalled breadth of choice, in terms of the routes and options we cover,” explains Michi Kossowsky, CTO at Freightos.com. “Instead of going through a freight forwarder, retailers and distributors can easily search for spot shipping quotes and compare them by price, estimated transit time, and so on. They can then track their shipments once they are en route.” As well as getting the best price, the user can also enjoy a more frictionless experience.
Because the Freightos.com platform relies on processing large volumes of data very quickly, Freightos.com chose to build it on the cloud. Google Cloud stood out, as it offered a unique solution to its hosting needs for in-memory processing. Once the platform was up and running, Freightos.com wanted to speed up its testing and delivery cycle, as well as create a more efficient workflow for its international teams of developers. Moving from a monolithic infrastructure to microservices on Google Kubernetes Engine (GKE) was the answer.
“Our goal at Freightos.com is to help people work together across the globe,” says Michi. “To do that, we wanted to work with the best hosting and mapping provider. Along with Google Maps Platform and G Suite, Google Cloud is at the heart of what we do at Freightos.com.”
While many companies now work with large datasets, Freightos.com’s challenge was exceptional. “We receive hundreds of thousands of searches a day, and to generate quotes, we search through tens of millions of data points supplied by more than 75 freight providers,” explains Michi. “So to run the graph-search algorithms, our app needs access to high-speed in-memory data.”
Freightos.com found a solution to this challenge by using App Engine flexible environments, providing access to 200 GB of routing data. However, it realized that breaking up its monolithic services would enable its development teams to work it a much more agile way.
To do that, Freightos.com worked with implementation partner DoiT International to migrate to GKE. “DoiT International has been providing Freightos.com with consulting services and support from the beginning,” explains Tony Braun, Cloud and Geolocation Sales Manager at DoiT International. “We also support Freightos.com in using Google Maps Platform, as its software depends on Maps geo-location, routing, and place information APIs.”
Freightos.com took a multiphased approach, re-architecting its infrastructure before making the switch. Freightos.com now uses GKE with Compute Engine, Pub/Sub, and Dataflow for its application pipeline. Cloud SQL and Datastore provide operational storage, while BigQuery and Google Data Studio offer data warehousing and data visualization.
“BigQuery is the bread and butter of our analytics, enabling us to query tremendous amounts of data very quickly, with little maintenance,” says Michi. “It has helped us to do things such as put together a global container freight index. Thanks to BigQuery, we can update the index every day.”
Freightos.com also offers its platform software on a SaaS basis, so freight forwarders can quickly and easily provide shipping rate quotes to direct customers. As part of this offering, Freightos.com uses the Apigee API Management Platform with Apigee developer portal to export APIs so they can be accessed by freight forwarders. “Apigee makes it easy to export and expose our APIs, encouraging adoption amongst our customers,” says Michi.
Coordinating teams across borders
Since switching to microservices using GKE, Freightos.com is finding it easier to coordinate its development teams and increase efficiency. That’s important, because it has teams based in Barcelona, Jerusalem, and Ramallah.
“The ability to work together and collaborate closely has been transformative,” says Michi. “Working with our other development teams on projects is one of my favourite things about working at Freightos.com. It is helping to change perceptions about what is possible, in terms of working together across geopolitical divides.”
G Suite is also supporting Freightos.com’s teams in collaborating, whether that be editing in real time using Docs, or holding a meeting using Hangouts Meet. “It’s central to the way we work,” says Michi. “Almost everything that happens in the company is facilitated by Google tools.”
Previously, Freightos.com’s development team had to follow a number of steps to test code before moving into production, which was time consuming. Now, it has a CI/CD pipeline that has automated that process, as well as enabled teams to work independently on individual services. “Thanks to Google Kubernetes Engine, it’s easier to scale up or down, and because the testing and production environments are the same, it’s much easier to deploy,” says Michi. “Teams can work more autonomously, and our platform is more stable as a result.”
Michi also appreciates the operations tools that Google Cloud offers for monitoring, such as Cloud Monitoring. “DoiT International has been really great in terms of helping us see where we can be more efficient,” says Michi. “We use its reOptimize platform for monthly cost predictions and to continuously optimize our Google Cloud activity.”
Freightos.com is now looking into possible applications for Google Cloud machine learning tools. “Some interesting applications for us could include using machine learning to automatically identify and match commodities to HS (product classification) codes, creating a predictive transit time estimator based on past carrier performance, and analyzing seasonal and weather factors,” says Michi.
It hopes that it can act as a trailblazer in helping people to work with one another. “Our goal is not only to be the number one freight marketplace, but to act as a real hub, driving standardization across the industry and helping connect all the different parties,” says Michi. “The Google services we use are fundamental to that vision.”