Client Case: appflame

How appflame's Team Efficiently Migrates Infrastructure to the Cloud with Google Cloud Solutions

About the company: appflame is a product-focused IT company that creates Ukrainian products that change lives and redefine the game’s rules in many categories. The company is based in Ukraine but operates globally. appflame has assembled a bold team of individuals who are passionate about creating real technological products and are determined to move forward, overcoming any obstacles in their path.

Company Goals:

1. To rank among the top 5 product IT companies in the world.
2. To become a unicorn.
3. To promote Ukraine with high-quality products in international markets.

Project Start Date: December 29, 2022

Project End Date: March 29, 2023

Challenge: The majority of appflame’s infrastructure was hosted on local servers. While the company had been using Firebase, a mobile app development platform by Google Cloud, for some time, they started actively using Google Cloud Platform (GCP) services towards the end of 2020 and the beginning of 2021. Their primary goal was to transition a significant portion of their infrastructure from local servers to the cloud and obtain a reliable, understandable, and continuously evolving cloud service provider, which Google Cloud had eventually become for the company.

Solution: Leveraging the wide array of Google Cloud tools, the appflame team decided to fulfill their other requirements using solutions from this cloud provider. They initiated a partnership with Cloudfresh, a Google Cloud Premier Partner, who provided appflame with valuable consultations regarding GCP services and helped select the most relevant solutions for their specific needs. After technical consultations and assessments, Cloudfresh devised a migration plan from local servers to the cloud to facilitate a seamless transition of the required infrastructure.

The appflame team created a template with Kubernetes cluster configurations, Cloud SQL, and Memorystore in Terraform, an infrastructure-as-code tool, enabling safe and efficient creation and modification of cloud and on-premises resource versions. Utilizing these solutions became critically important for appflame’s products. After configuration, they achieved:

• Kubernetes (k8s): Balanced orchestration of all containers, serving as the primary deployment unit for applications in k8s. With this service, the company could horizontally scale applications by adding or removing containers based on workload, resulting in a fully balanced traffic distribution between different container instances for optimal resource utilization.

• Cloud SQL: A fully managed relational database service for MySQL, PostgreSQL, and SQL Server, which freed the team from database administration tasks. By employing various CloudSQL services such as backups, network connectivity, monitoring, and more, appflame no longer needed to create and maintain databases themselves but rather managed and interacted with them.

• Memorystore: An in-memory data storage service for Redis Cluster, Redis, and Memcached. This service enabled the team to create application caches, provide data access in milliseconds, and move the caching layer to the cloud without any code changes, with simplified scaling without downtime.

The appflame team also optimized their resources by utilizing Spot Virtual Machines (Spot nodes) following Google’s best practices. Compared to standard machines, they are significantly cheaper and utilize the Compute Engine’s excess capacity, making their availability dependent on usage. Compute Engine may stop these machines if it needs to reclaim computational power for distributing other virtual machines. However, with the use of necessary automation tools for specific workloads, this solution is highly cost-effective and allows deploying required capacities with significantly lower costs. The company is also fully satisfied with the PCI DSS Compliance coverage for their entire infrastructure, ensuring data security.

Results: Through a fruitful collaboration with Cloudfresh, appflame successfully migrated a portion of their infrastructure related to databases from local servers to the cloud and started utilizing cloud capabilities more intensively after the partial migration.

Over a relatively short period, the company achieved significant outcomes in optimizing their own infrastructure:

1. Leveraging services like Cloud SQL and Memory Store significantly improved their database management, enabling employees to optimize their time resources and focus on more strategic tasks.
2. Kubernetes provided the team with a convenient framework for the flexible operation of distributed systems. It handles scalability error handling in applications, provides deployment templates, and much more.
3. Spot virtual machines contributed to substantial resource savings. Now, the company uses GCP resources at a discount of 60-90%, and teams process data in an automated manner, swiftly deploying the necessary infrastructure in any region.

Cloudfresh’s Role: The Cloudfresh team helped streamline all of appflame’s resources, provided technical consulting, and assisted in selecting GCP services relevant to the company’s needs.

Certified technical specialists at Cloudfresh developed an action plan for migrating from a bare-metal provider to Google Cloud, offering technical consultations on architecture and services tailored to the company’s requirements. Furthermore, Cloudfresh helped appflame significantly reduce resource costs through special conditions, discounts, and a convenient payment method.

The teams continue to collaborate to this day. Cloudfresh experts happily provide ongoing consultations to appflame throughout the migration process and support the team on their journey toward transitioning infrastructure to the cloud.