Transactions across serverless functions leveraging stateful dataflows

Serverless computing is currently the fastest-growing cloud services segment. The most prominent serverless offering Function-as-a-Service (FaaS), where users write functions and automates deployment, maintenance, scalability. Although FaaS a good fit for executing stateless functions, it does not adequately support stateful constructs like microservices scalable, low-latency applications. Recently, there have been multiple attempts to add first-class state in systems, such as Microsoft Orleans, Azure Durable Functions, or Beldi. These approaches execute business code inside handing over their an external database. In contrast, Apache Flink’s StateFun follow different design: dataflow system Flink handles all management, messaging, checkpointing by graph providing exactly-once processing guarantees. This design relieves programmers from having “pollute” logic with distributed systems error checking, mitigation. can easily develop applications without worrying about messaging transactions across remains open problem. this paper, we introduce programming model implementation transaction orchestration of functions. Our supports serializable two-phase commit, well eventually consistent workflows Sagas. We implement our on leverage management experiments show that approach building transactional top graphs achieve very high throughput, but latency overhead due mechanism guaranteeing processing. compare Beldi implements commit AWS lambda backed DynamoDB makes use CockroachDB its backend.