Aiven helps Paf stay ahead with Apache Kafka^®

Overview

Paf is a physical and online gambling operator in the Baltics run by the Government of Åland. They are moving from classic mechanical games to modern online hybrid gaming, and because of their mix of legacy and current technologies, they need Aiven for Apache Kafka® to do it.

Highlights

• Paf's infrastructure is a mix of legacy and current systems that Aiven for Apache Kafka helps them juggle and streamline.
• The infra moves around data from data warehousing, Business Intelligence and CRM applications.
• Handling identifying information means GDPR compliance both at rest and in transit.

The challenge

Paf is a physical and online gambling operator in the Baltics run by the Government of Åland with a mission to generate profit for the benefit of the Ålandic society.

Many of Paf’s data pipelines leverage a mix of legacy and current technologies, which requires building loosely-coupled applications and systems to support data schemas. Moreover, the implementation needs to work in real time.

Several disparate teams at the company – with more than 350 employees at last count – have had to create a variety of data pipelines to support a range of existing and yet-unknown future systems.

Robarto Zagni, Paf’s Head of Data and CRM, explains the difficulty: “Our data warehousing, BI, and CRM applications provide marketing and support teams with the data they need. In that function, we are also the ones building infrastructure to move the data around...and while [our old] systems work, they’re not so open or easily interfaceable for other parts of the company.”

Additionally, Paf needs to be able to quickly handle GDPR - particularly related to identifying information - in data that is both in transit and at rest. A system like Apache Kafka enables this - but how to begin?

The solution

Zagni reported that he viewed running Kafka in-house as a risk primarily because no one in the company had experience in running Kafka clusters. They could either train or hire people, but these were both inefficient routes.

So Paf looked into a third option: buying Kafka as a service. After considering Confluent’s hosted Kafka, they decided to let Aiven handle the SLA, monitor the cluster, and maintain security updates.

Besides competitive pricing, the flexibility of future schemas and the freedom from data setup and management overhead were key factors in deciding to go with Aiven for Kafka. Manager of System Ops, Fredrik Wiklund explains:

“Aiven for Kafka was a really good option for us because, as we’re moving to AWS, we could use the Bring Your Own Account option. We handle the AWS accounts and have more control over our data, while Aiven handles Kafka.”

Wiklund also enjoys the human touch and dedication to security that working with the team entails:

“We can just walk over to them if we have anything we want to discuss. And Aiven is ISO 27001 certified already. We did an external audit on Aiven’s Information Security Management System and were really impressed about how Aiven is taking information security seriously.”

The outcome

Paf had multiple teams serving a myriad of purposes, not always a clear interface between many functions (and data pipelines), the need to spin up custom pipelines fast, and the scrutiny of both local regulatory requirements and GDPR. They sorely needed a messaging solution that was secure, loosely-coupled, flexible, transparent, and managed, so that they could stay focused on their core business. At the intersection of all of these is Aiven for Apache Kafka®.