Strome: Energy-Aware Data-Stream Processing

  title={Strome: Energy-Aware Data-Stream Processing},
  author={Christopher Eibel and Christian Gulden and Wolfgang Schr{\"o}der-Preikschat and Tobias Distler},
Handling workloads generated by a large number of users, data-stream–processing systems also require large amounts of energy. To reduce their energy footprint, such systems typically rely on the operating systems of their servers to adjust processor speeds depending on the current workload by performing dynamic voltage and frequency scaling (DVFS). In this paper, we show that, although effective, this approach still leaves room for significant energy savings due to DVFS making conservative… 
Run-time Adaptation of Data Stream Processing Systems: The State of the Art
This survey reviews the most relevant approaches from the literature, presenting a taxonomy to characterize the state of the art along several key dimensions and allows us to identify current research trends as well as open challenges that will motivate further investigations in this field.
Stream-based State-Machine Replication
This paper shows how stream-based replication, a novel approach that implements a replication protocol as application on top of a data-stream processing framework, has the key benefit of significantly minimizing overhead for both programmers as well as system operators.


Empya: Saving Energy in the Face of Varying Workloads
Empya is an energy-aware programming and execution platform that frees application programmers from the need to take care of energy efficiency and combines techniques from different software and hardware levels to effectively and efficiently minimize the resource footprint of an application during periods of low utilization.
Towards energy proportionality for large-scale latency-critical workloads
PEGASUS is presented, a feedback-based controller that significantly improves the energy proportionality of WSC systems, as demonstrated by a real implementation in a Google search cluster.
Keep calm and react with foresight: strategies for low-latency and energy-efficient elastic data stream processing
The results demonstrate the high-degree of flexibility and configurability of the approach, and show the effectiveness of the elastic scaling strategies compared with existing state-of-the-art techniques used in similar scenarios.
Adaptive Provisioning of Stream Processing Systems in the Cloud
This paper proposes an adaptive approach for provisioning virtual machines for the use of a DSPS in the cloud that reacts to changes in the stream workload and shows that this approach can achieve low-latency stream processing when VMs are not overloaded, while adjusting resources dynamically with workload changes.
Online parameter optimization for elastic data stream processing
This paper presents an elastic scaling data stream processing prototype, which allows to trade off monetary cost against the offered quality of service and uses an online parameter optimization, which minimizes the monetary cost for the user.
StreamCloud: An Elastic and Scalable Data Streaming System
StreamCloud is presented, a scalable and elastic stream processing engine for processing large data stream volumes that uses a novel parallelization technique that splits queries into subqueries that are allocated to independent sets of nodes in a way that minimizes the distribution overhead.
Elastic stateful stream processing in storm
This paper extends Storm with new components that allow automatic elasticity and stateful migration of the application components and shows the benefits of the newly introduced functionalities that allow to properly cope with workload variations while improving the resource utilization of the underlying infrastructure.
Towards Energy-Proportional State-Machine Replication
The evaluation results for a Byzantine fault-tolerant coordination service show that utilizing such knowledge in combination with the mechanism presented, it is possible to build energy-proportional replicated systems.
Workload analysis of a large-scale key-value store
This paper collects detailed traces from Facebook's Memcached deployment, arguably the world's largest, and analyzes the workloads from multiple angles, including: request composition, size, and rate; cache efficacy; temporal patterns; and application use cases.
Beyond DVFS: A First Look at Performance under a Hardware-Enforced Power Bound
A first look at dynamic Voltage Frequency Scaling in the HPC environment is provided and it is shown that execution under a power bound translates this variation in efficiency into variation in performance.