Confluent, Inc. announced AI Model Inference, an upcoming feature on Confluent Cloud for Apache Flink, to enable teams to easily incorporate machine learning into data pipelines. Confluent introduced Confluent Platform for Apache Flink®, a Flink distribution that enables stream processing in on-premises or hybrid environments with support from the company's Flink experts. Confluent also unveiled Freight clusters, a new cluster type for Confluent Cloud that provides a cost-effective way to handle large-volume use cases that aren't time-sensitive, such as logging or telemetry data.

AI Model Inference simplifies building and launching AI and ML applications: Generative AI helps organizations innovate faster and deliver more tailored customer experiences. AI workloads need fresh, context-rich data to ensure the underlying models generate accurate output and results for businesses to make informed decisions based on the most current information available. However, developers often have to use several tools and languages to work with AI models and data processing pipelines, leading to complex and fragmented workloads.

This can make it challenging to leverage the most current and relevant data for decision-making, leading to errors or inconsistencies and compromising the accuracy and reliability of AI-driven insights. These issues can cause increased development time and difficulties in maintaining and scaling AI applications. With AI Model Inference in Confluent Cloud for Apache Flink®, organizations can use simple SQL statements from within Apache Flink to make calls to AI engines, including OpenAI, Amazon SageMaker, GCP Vertex, and Microsoft Azure.

Now enterprises can orchestrate data cleaning and processing tasks on a single platform. AI Model Inference enables companies to: Simplify AI development by using familiar SQL syntax to work directly with AI/ML models, reducing the need for specialized tools and languages; Establish seamless coordination between data processing and AI workflows to improve efficiency and reduce operational complexity; Enable accurate, real-time AI-driven decision-making by leveraging fresh, contextual streaming data. Confluent Platform for Apache Flink® enables stream processing in private clouds and on-premises environments: Many organizations are looking for hybrid solutions to protect more sensitive workloads.

With Confluent Platform for Apache Flink®, a Flink distribution fully supported by Confluent, customers can easily leverage stream processing for on-prem or private cloud workloads with long-term expert support. Apache Flink can be used in tandem with Confluent Platform with minimal changes to existing Flink jobs and architecture. Confluent Platform for Apache Flink® can help organizations: Minimize risk with unified Flink and Kafka support and expert guidance from the foremost experts in the data streaming industry; Receive timely assistance in troubleshooting and resolving issues, reducing the impact of any operational disruptions to mission critical applications; Ensure that stream processing applications are secure and up to date with off-cycle bug and vulnerability fixes.

With Kafka and Flink available in Confluent?s complete data streaming platform, organizations can ensure better integration and compatibility between technologies, and receive comprehensive support for streaming workloads across all environments. Unlike open source Apache Flink which only maintains the two most recent releases, Confluent offers three years of support for every Confluent Platform for Apache Flink® release from launch, guaranteeing uninterrupted operations and peace of mind. Confluent Platform for Apache Flink® will be available to Confluent customers later 2024.

New auto-scaling Freight clusters offer more cost-efficiency at scale: Many organizations use Confluent Cloud to process logging and telemetry data. These use cases involve large amounts of business-critical data but are often less latency-sensitive since they typically feed into indexing or batch aggregation engines. To make these common use cases more cost-efficient for customers, Confluent is introducing Freight clusters?a new serverless cluster type with up to 90% lower cost for high-throughput use cases with relaxed latency requirements.

Powered by Elastic CKUs, Freight clusters seamlessly auto-scale based on demand with no manual sizing or capacity planning required, enabling organizations to minimize operational overhead and optimize costs by paying only for the resources that they use when they need it. Freight clusters are available in early access in select AWS regions.