Sören Henning

Sören Henning, Otmar Ertl, Adriano Vogel

Shuffling or repartitioning data streams is an essential operation of state-of-the-art stream processing frameworks to support stateful workloads in a large-scale, distributed setting. In today's cloud deployments, however, shuffling can become a major cost driver due to substantial network traffic across multiple availability zones (AZs) as well as an operational burden when operating a high-throughput, strongly consistent messaging backbone at scale. We present BlobShuffle, a novel approach to cost-effective shuffling for stream processing systems that leverages cloud object storage as an intermediate exchange layer. Instead of sending all shuffled records directly, BlobShuffle groups records into batches, stores these batches in cloud object storage, and forwards only compact notifications. Downstream operators use these notifications to retrieve the relevant batches and extract the corresponding records. BlobShuffle balances cost efficiency and latency through configurable batching and a distributed caching mechanism. BlobShuffle is implemented as an add-on for Kafka Streams that requires only minimal code changes to existing applications, leaves Kafka and the underlying infrastructure unmodified, and preserves Kafka Streams' consistency and correctness guarantees. In a large-scale experimental evaluation on a Kubernetes-based AWS deployment, we show that BlobShuffle can reduce shuffling costs by more than 40x compared to native Kafka Streams shuffling while keeping the 95th percentile shuffle latency below 2 seconds. Moreover, it scales to processing more than 2 GiB/s without encountering a scalability limit in our experiments, indicating that BlobShuffle can economically support shuffle-intensive workloads at large scale.

Sören Henning, Wilhelm Hasselbring

Distributed stream processing engines are designed with a focus on scalability to process big data volumes in a continuous manner. We present the Theodolite method for benchmarking the scalability of distributed stream processing engines. Core of this method is the definition of use cases that microservices implementing stream processing have to fulfill. For each use case, our method identifies relevant workload dimensions that might affect the scalability of a use case. We propose to design one benchmark per use case and relevant workload dimension. We present a general benchmarking framework, which can be applied to execute the individual benchmarks for a given use case and workload dimension. Our framework executes an implementation of the use case's dataflow architecture for different workloads of the given dimension and various numbers of processing instances. This way, it identifies how resources demand evolves with increasing workloads. Within the scope of this paper, we present 4 identified use cases, derived from processing Industrial Internet of Things data, and 7 corresponding workload dimensions. We provide implementations of 4 benchmarks with Kafka Streams and Apache Flink as well as an implementation of our benchmarking framework to execute scalability benchmarks in cloud environments. We use both for evaluating the Theodolite method and for benchmarking Kafka Streams' and Flink's scalability for different deployment options.

Sören Henning, Wilhelm Hasselbring

Ever-increasing amounts of data and requirements to process them in real time lead to more and more analytics platforms and software systems being designed according to the concept of stream processing. A common area of application is the processing of continuous data streams from sensors, for example, IoT devices or performance monitoring tools. In addition to analyzing pure sensor data, analyses of data for groups of sensors often need to be performed as well. Therefore, data streams of the individual sensors have to be continuously aggregated to a data stream for a group. Motivated by a real-world application scenario, we propose that such a stream aggregation approach has to allow for aggregating sensors in hierarchical groups, support multiple such hierarchies in parallel, provide reconfiguration at runtime, and preserve the scalability and reliability qualities induced by applying stream processing techniques. We propose a stream processing architecture fulfilling these requirements, which can be integrated into existing big data architectures. We present a pilot implementation of such an extended architecture and show how it is used in industry. Furthermore, in experimental evaluations we show that our solution scales linearly with the amount of sensors and provides adequate reliability in the case of faults.