Event streaming can be transformative for businesses, with numerous use cases across all industry sectors.
This emerged during a webinar hosted by Confluent, in partnership with ITWeb, in which Dan Croft, senior solutions engineer at Confluent, outlined use cases and examples of how event streaming with Apache Kafka and Confluent has made organisations more efficient and competitive.
“Organisations all face the same pressures – their customer expectations are going up exponentially. They want more done, faster and cheaper.”
Improved service and customer satisfaction are key across everything from car manufacturers to telcos, he said.
“To create those real-time customer relationships, companies need real time streams. They are looking to Apache Kafka to help them.”
Croft explained: “Apache Kafka is not another database or data store. We need to feed real-time information as this information is only really valuable in real time. Every time there’s a new update on a database, it pushes through into Kafka, and people can pull it through Kafka instead of from the database. Under the covers, it’s very simple: it groups individual packets or messages into topics, and anyone can become a subscriber to that topic, seeing only the latest updates or going back in time. All of this data is kept, and is immutable.”
How big can we make Kafka? LinkedIn uses it at a massive scale. It’s the perfect use case for Kafka.
Dan Croft, Confluent.
Kafka has a huge amount of horsepower to bring to bear. It will sit on multiple systems and deliver multiple throughputs, he said. “How big can we make Kafka? LinkedIn uses it at a massive scale. It’s the perfect use case for Kafka – a single update in your feed has to be distributed to hundreds of people. My LinkedIn update goes into a Kafka topic, and my contacts – or consumers – will see that update. That happens all day, every day, with 800 million members on the LinkedIn network.”
Confluent brings the enterprise version of Kafka®, contributing around 80% of the code to the Apache Project, and adding tooling and support across the Confluent Platform on-premise solution and the Confluent Cloud SaaS version.
Confluent highlights some of the top use cases for Kafka as customer 360° and website clickstream analysis, legacy IT modernisation, next-generation apps, achieving a single business view, and real-time analytics.
Croft outlined a number of case studies in which large enterprises had transformed areas of their business with Apache Kafka and Confluent.
“Nationwide Building Society uses Confluent Platform to increase digital activations and improve customer experience. With Audi, we worked with an automation solution on the AWS backbone delivering key statistics such as interruptions, false starts, noises or vibrations in the engine. Audi deployed Automotive Cloud Data Collector – a streaming platform powered by Confluent – to enrich its connected car driving experience, with real-time data streamed from vehicle sensors.”
Confluent also worked with ao.com, a retailer of larger appliances who needed to keep historical and browsing data to better serve their customers with a hyper personalised approach. As a result, the company achieved a 30% improvement in conversion rates.
Expedia used Apache Kafka and Confluent Cloud to build a fast, super-scalable event driven architecture that could handle immense traffic spikes and also provide guarantees around delivery semantics. Confluent Cloud for them was ideal, because it is elastic, global and has infinite storage. Expedia can now minimise capacity planning and eliminate cluster management, and handle massive traffic spikes without lag or downtime.
“For financial services, it’s an easy way to commoditise data. Fintech Q2 uses Confluent Platform and Apache Kafka to implement a scalable event-driven architecture that powers the alerting system and broader digital transformation, while assuring security and compliance,” Croft said.
Confluent notes that while event streaming can revolutionise business, to truly take advantage of it as a strategic platform, it must move from the domain of one-off engineering initiatives to the enterprise’s central nervous system.
Share