Refine
Document Type
- Conference proceeding (9)
- Journal article (6)
Language
- English (15)
Is part of the Bibliography
- yes (15)
Institute
- Informatik (15)
Publisher
- Springer (8)
- SciTePress (3)
- IEEE (2)
- IBM Research Division (1)
- World Scientific Publishing (1)
Serverless computing is an emerging cloud computing paradigm with the goal of freeing developers from resource management issues. As of today, serverless computing platforms are mainly used to process computations triggered by events or user requests that can be executed independently of each other. These workloads benefit from on-demand and elastic compute resources as well as per-function billing. However, it is still an open research question to which extent parallel applications, which comprise most often complex coordination and communication patterns, can benefit from serverless computing.
In this paper, we introduce serverless skeletons for parallel cloud programming to free developers from both parallelism and resource management issues. In particular, we investigate on the well known and widely used farm skeleton, which supports the implementation of a wide range of applications. To evaluate our concepts, we present a prototypical development and runtime framework and implement two applications based on our framework: Numerical integration and hyperparameter optimization - a commonly applied technique in machine learning. We report on performance measurements for both applications and discuss
the usefulness of our approach.
Elasticity is considered to be the most beneficial characteristic of cloud environments, which distinguishes the cloud from clusters and grids. Whereas elasticity has become mainstream for web-based, interactive applications, it is still a major research challenge how to leverage elasticity for applications from the high-performance computing (HPC) domain, which heavily rely on efficient parallel processing techniques. In this work, we specifically address the challenges of elasticity for parallel tree search applications. Well-known meta-algorithms based on this parallel processing technique include branch-and-bound and backtracking search. We show that their characteristics render static resource provisioning inappropriate and the capability of elastic scaling desirable. Moreover, we discuss how to construct an elasticity controller that reasons about the scaling behavior of a parallel system at runtime and dynamically adapts the number of processing units according to user-defined cost and efficiency thresholds. We evaluate a prototypical elasticity controller based on our findings by employing several benchmarks for parallel tree search and discuss the applicability of the proposed approach. Our experimental results show that, by means of elastic scaling, the performance can be controlled according to user-defined thresholds, which cannot be achieved with static resource provisioning.
In recent years, the cloud has become an attractive execution environment for parallel applications, which introduces novel opportunities for versatile optimizations. Particularly promising in this context is the elasticity characteristic of cloud environments. While elasticity is well established for client-server applications, it is a fundamentally new concept for parallel applications. However, existing elasticity mechanisms for client-server applications can be applied to parallel applications only to a limited extent. Efficient exploitation of elasticity for parallel applications requires novel mechanisms that take into account the particular runtime characteristics and resource requirements of this application type. To tackle this issue, we propose an elasticity description language. This language facilitates users to define elasticity policies, which specify the elasticity behavior at both cloud infrastructure level and application level. Elasticity at the application level is supported by an adequate programming and execution model, as well as abstractions that comply with the dynamic availability of resources. We present the underlying concepts and mechanisms, as well as the architecture and a prototypical implementation. Furthermore, we illustrate the capabilities of our approach through real-world scenarios.
Cloud resources can be dynamically provisioned according to application-specific requirements and are payed on a per-use basis. This gives rise to a new concept for parallel processing: Elastic parallel computations. However, it is still an open research question to which extent parallel applications can benefit from elastic scaling, which requires resource adaptation at runtime and corresponding coordination mechanisms. In this work, we analyze how to address these system-level challenges in the context of developing and operating elastic parallel tree search applications. Based on our findings, we discuss the design and implementation of TASKWORK, a cloud-aware runtime system specifically designed for elastic parallel tree search, which enables the implementation of elastic applications by means of higher-level development frameworks. We show how to implement an elastic parallel branch-and-bound application based on an exemplary development framework and report on our experimental evaluation that also considers several benchmarks for parallel tree search.
The cloud evolved into an attractive execution environment for parallel applications, which make use of compute resources to speed up the computation of large problems in science and industry. Whereas Infrastructure as a Service (IaaS) offerings have been commonly employed, more recently, serverless computing emerged as a novel cloud computing paradigm with the goal of freeing developers from resource management issues. However, as of today, serverless computing platforms are mainly used to process computations triggered by events or user requests that can be executed independently of each other and benefit from on-demand and elastic compute resources as well as per-function billing. In this work, we discuss how to employ serverless computing platforms to operate parallel applications. We specifically focus on the class of parallel task farming applications and introduce a novel approach to free developers from both parallelism and resource management issues. Our approach includes a proactive elasticity controller that adapts the physical parallelism per application run according to user-defined goals. Specifically, we show how to consider a user-defined execution time limit after which the result of the computation needs to be present while minimizing the associated monetary costs. To evaluate our concepts, we present a prototypical elastic parallel system architecture for self-tuning serverless task farming and implement two applications based on our framework. Moreover, we report on performance measurements for both applications as well as the prediction accuracy of the proposed proactive elasticity control mechanism and discuss our key findings.