Data Stream Operations as First-Class Entities in Component-Based Performance Models

Abstract

Data streaming applications are an important class of dataintensive systems. Performance is an essential quality of such systems. It is, for example, expressed by the delay of analysis results or the utilization of system resources. Architecture-level decisions such as the configuration of sources, sinks and operations, their deployment or the choice of technology impact the performance. Current component-based performance prediction approaches cannot accurately predict the performance of those systems, because they do not support the metrics that are specific to data streaming applications and only approximate the behavior of data stream operations instead of expressing it explicitly. In particular, operations that group multiple data events and thus introduce timing dependencies between different calls to the system are not represented sufficiently. In this paper, we present an approach for modeling networks of data stream operations including their parameters with the goal of predicting the performance of the resulting composed data streaming application. The approach is based on a component-based performance model with queueing semantics for processing resources. Our evaluation shows that our model can more accurately express the behavior of the system, resulting in a more expressive performance model compared to a well-encapsulated component-based model without data stream operations.

A Unified Model to Detect Information Flow and Access Control Violations in Software Architectures

“…Allocations assign component instances to resource nodes. The Palladio extension for data-oriented modeling (Werle et al, 2020) introduces data channels as special components that communicate by emitting and consuming data items instead of doing calls.…”

“…We realize our concepts within the ADL Palladio (Reussner et al, 2016) in order to identify the required steps to extend existing ADLs. We chose Palladio for two reasons: First, Palladio is capable of describing behavior in terms of control flow and data flow (Werle et al, 2020), so we can evaluate our concepts for both paradigms within one prototype. Second, Palladio provides a condensed set of modeling primitives tailored for analyzing architectural quality properties, which allows us to focus on core concepts.…”

Accurate Performance Predictions with Component-Based Models of Data Streaming Applications

“…We use a case study to evaluate the benefit of models extended with SAOs. We extend our previous research in performance modeling for data-streaming applications [19] with a generalization of the model and simulation and with an evaluation of the resulting approach.…”

“…We call these operations stateful asynchronous operations (SAOs). This question is relevant because our previous research [19] indicates that some of the timing effects present in this type of system cannot be suitably predicted with stateless models and require approximations and workarounds. An example of a stateful operation is when data is collected and emitted as a group after a specified duration has passed or after a specified number of elements has arrived.…”

“…1. Illustration of the case study system (Source: [18]) calls within the system, and b) it is not as well supported with current modeling approaches. Our evaluation shows that MSAO yields prediction results that are competitive with BL without incurring the additional manual engineering overhead.…”