Most distributed stream processing engines (DSPEs) do not support online task management and cannot adapt to time-varying data flows. Recently, some studies have proposed online task deployment algorithms to solve this problem. However, these approaches do not guarantee the Quality of Service (QoS) when the task deployment changes at runtime, because the task migrations caused by the change of task deployments will impose an exorbitant cost. We study one of the most popular DSPEs, Apache Storm, and find out that when a task needs to be migrated, Storm has to stop the resource (implemented as a process of Worker in Storm) where the task is deployed. This will lead to the stop and restart of all tasks in the resource, resulting in the poor performance of task migrations. Aiming to solve this problem, in this paper, we propose N-Storm (Nonstop Storm), which is a task-resource decoupling DSPE. N-Storm allows tasks allocated to resources to be changed at runtime, which is implemented by a thread-level scheme for task migrations. Particularly, we add a local shared key/value store on each node to make resources aware of the changes in the allocation plan. Thus, each resource can manage its tasks at runtime. Based on N-Storm, we further propose Online Task Deployment (OTD). Differing from traditional task deployment algorithms that deploy all tasks at once without considering the cost of task migrations caused by a task re-deployment, OTD can gradually adjust the current task deployment to an optimized one based on the communication cost and the runtime states of resources. We demonstrate that OTD can adapt to different kinds of applications including computation- and communication-intensive applications. The experimental results on a real DSPE cluster show that N-Storm can avoid the system stop and save up to 87% of the performance degradation time, compared with Apache Storm and other state-of-the-art approaches. In addition, OTD can increase the average CPU usage by 51% for computation-intensive applications and reduce network communication costs by 88% for communication-intensive applications.

The recently proposed learned index has higher query performance and space efficiency than the conventional B+-tree. However, the original learned index has the problems of insertion failure and unbounded query complexity, meaning that it supports neither insertions nor bounded query complexity. Some variants of the learned index use an out-of-place strategy and a bottom-up build strategy to accelerate insertions and support bounded query complexity, but introduce additional query costs and frequent node splitting operations. Moreover, none of the existing learned indices are cache-friendly. In this paper, aiming to not only support efficient queries and insertions but also offer bounded query complexity, we propose a new learned index called COLIN (Cache-cOnscious Learned INdex). Unlike previous solutions using an out-of-place strategy, COLIN adopts an in-place approach to support insertions and reserves some empty slots in a node to optimize the node’s data placement. In particular, through model-based data placement and cache-conscious data layout, COLIN decouples the local-search boundary from the maximum error of the model. The experimental results on five workloads and three datasets show that COLIN achieves the best read/write performance among all compared indices and outperforms the second best index by 18.4%, 6.2%, and 32.9% on the three datasets, respectively.