Google Cloud Platform for Kaptio’s Travel Tech

Author: Peter Bridger

Spoiler Alert: this blog is for all the travel tech geeks out there and professionals wanting to understand how Kaptio is on the bleeding edge of technology.

Kaptio has always been a cloud-first organisation with our travel reservation platform and travel CRM built upon Salesforce, allowing us to provide a native SaaS (software as a service) solution to clients worldwide.

Salesforce provides a tremendous and ever-growing framework that we take advantage of to continuously improve and extend our product. However, to see Kaptio as purely, a team of Salesforce focused experts wouldn’t be the whole story. Read on to find out how we are continuing to improve our infrastructure.

Travel Tech Background

Over the years, we’ve migrated core parts of business logic and gateways to 3rd party partners out of Salesforce into a series of API based services that Salesforce can then call on to retrieve data or make live inventory bookings.

This approach has enabled us to take advantage of other platforms and languages to provide better scalability and ease of development in these areas. It’s also enabled us to maintain separate teams within Kaptio with ownership of and specialism in these separate domains.

Our Travel Reservation Platform’s Future

For much of our journey, we’ve made use of Heroku to provide cloud hosting for our API systems. Heroku sits on top of AWS (Amazon Web Services) whilst offering tooling to make it straightforward for developers to quickly and easily deploy and manage systems effectively in the cloud.

As Kaptio continues to grow, we’ve found ourselves hitting some of the limitations of Heroku and have explored other options.

After looking at the big three providers of AWS, Azure and GCP (Google Cloud Platform) and evaluating our requirements against their current and future offerings, we decided on a migration to GCP.

GCP graphic v4 600x300.png

Google Cloud Platform for Kaptio’s Travel CRM

It’s fair to say that Google has strong experience with cloud technologies, although that hasn’t historically translated into them gaining the same traction that AWS and Azure have enjoyed.

Recently GCP has begun to hit its stride, which has likely been helped by an industry-wide increase in adoption of containerisation and orchestration technologies such as Docker and Kubernetes. GCP provides a joined-up suite of services that complement each other and are backed with solid documentation.

Infrastructure as Code

GCP makes it quick and easy to spin up infrastructure through their web-based UI, along with providing an API that can easily be used via a local or web-based console. We wanted to ensure we had repeatability in how we provisioned our infrastructure between systems and environments beyond prototyping. In addition to having accountability through auditing to show when a change was made, by who and for what reason.

These desires quickly led us to Terraform by Hashicorp that provides an industry-leading tool for managing cloud infrastructure as code. Being a platform-agnostic tool, it can also just as easily work with AWS and Azure.

Travel Tech Architecture

GCP provides managed offerings for traditional services such as PostgreSQL through Cloud SQL or Memcached through Memorystore. This allows for instances of various sizes and levels of redundancy to be easily set up and reconfigured as our future needs dictate.

One of the bigger decisions we had to make was which of GCP’s compute services would best serve the needs for each of our systems.

GCP offers traditional VMs (Virtual Machines) that provide a very familiar environment when compared to traditional onsite hosting. However, the real benefits of GCP are realised through their newer hosting models, which enable more flexibility in terms of scalability and how costs can be managed.

With our focus on maintaining a series of independently scalable microservices, we explored GCPs Cloud Run and GKE (Google Kubernetes Engine) offerings that are designed for containerised systems.

Although it’s possible to use GKE to wield a great deal of power over how a system runs, we found that many of our systems would be better served by using Cloud Run.

Cloud Run makes it straightforward to host a request-driven containerised system, allowing instance sizing and scaling to be easily managed whilst taking care of network configuration and logging integration.

It also allows us to choose from a series of different deployment strategies, such as blue/green or canary, depending on the system.

Tight integration with GCP services such as Cloud Logging and Secrets Manager has allowed us to take advantage of sophisticated logging and access control and audit logs for sensitive credentials whilst moving away from the assortment of individual providers we traditionally used.

Meet the brains behind this shift.

GCP team v4 600x300.png

Continued Growth

As we move to more sophisticated workloads in the future, we plan to prototype further with Knative and tools such as Spinnaker to allow us to confidently manage a high-quality continuous delivery pipeline.

GCP also allows us to make use of mature big data capabilities to monitor behaviours and drive customer recommendations through our existing data stores.

As Kaptio continues to grow and expand, we’ve got confidence that GCP will continue to support us on that journey.

About Kaptio

Only Kaptio Travel and Salesforce, the world’s no 1 CRM system, delivers a truly customer-centric approach. With your customers at the core of your system, you can understand their needs and react quickly, offering them the certainty they need in an uncertain world. Get in touch today to discover why brands like Rocky Mountaineer and Rhino Africa rely on our travel CRM and travel reservation platform to power their operations.