arXiv subjects

Josef Spillner

Publications and source records attributed to Josef Spillner.

10 recordsLinked to original sources

arXiv2021

Analysis and Improvement of Heterogeneous Hardware Support in Docker Images

Docker images are used to distribute and deploy cloud-native applications in containerised form. A container engine runs them with separated privileges according to namespaces. Recent studies have investigated security vulnerabilities and runtime characteristics of Docker images. In contrast, little is known about the extent of hardware-dependent features in them such as processor-specific trusted execution environments, graphics acceleration or extension boards. This problem can be generalised to missing knowledge about the extent of any hardware-bound instructions within the images that may require elevated privileges. We first conduct a systematic one-year evolution analysis of a sample of Docker images concerning their use of hardware-specific features. To improve the state of technology, we contribute novel tools to manage such images. Our heuristic hardware dependency detector and a hardware-aware Docker executor give early warnings upon missing dependencies instead of leading to silent or untimely failures. Our dataset and tools are released to the research community.

cs.CR↗

arXiv2020

Monitoring Data Distribution and Exploitation in a Global-Scale Microservice Artefact Observatory

Reusable microservice artefacts are often deployed as black or grey boxes, with little concern for their properties and quality, beyond a syntactical interface description. This leads application developers to chaotic and opportunistic assumptions about how a composite application will behave in the real world. Systematically analyzing and tracking these publicly available artefacts will grant much needed predictability to microservice-based deployments. By establishing a distributed observatory and knowledge base, it is possible to track microservice repositories and analyze the artefacts reliably, and provide insights on their properties and quality to developers and researchers alike. This position paper argues for a federated research infrastructure with consensus voting among participants to establish and preserve ground truth about the insights.

cs.SE↗

arXiv2019

Quantitative Analysis of Cloud Function Evolution in the AWS Serverless Application Repository

The serverless computing ecosystem is growing due to interest by software engineers. Beside Function-as-a-Service (FaaS) and Backend-as-a-Service (BaaS) systems, developer-oriented tools such as deployment and debugging frameworks as well as cloud function repositories enable the rapid creation of wholly or partially serverless applications. This study presents first insights into how cloud functions (Lambda functions) and composite serverless applications offered through the AWS Serverless Application Repository have evolved over the course of one year. Specifically, it outlines information on cloud function and function-based application offering models and descriptions, high-level implementation statistics, and evolution including change patterns over time. Several results are presented in live paper style, offering hyperlinks to continuously updated figures to follow the evolution after publication date.

cs.DC↗

arXiv2019

Quality Assessment and Improvement of Helm Charts for Kubernetes-Based Cloud Applications

Helm has recently been proposed by practitioners as technology to package and deploy complex software applications on top of Kubernetes-based cloud computing platforms. Despite growing popularity, little is known about the individual so-called Helm Charts and about the emerging ecosystem of charts around the KubeApps Hub website and decentralised charts repositories. This article contributes first quantified insights around both the charts and the artefact development community based on metrics automatically gathered by a proposed quality assessment tool named HelmQA. The work further identifies quality insufficiencies detectable in public charts, proposes a developer-centric hypothesis-based methodology to systematically improve the quality by using HelmQA, and finally empirically attempts to validate the methodology and thus the practical usefulness of the tool by presenting results of its application over a representative four-month period. Although one of our initial hypotheses does not statistically hold during the experiment, we still infer that using HelmQA regularly in continuous software development would lead to reduced quality issues.

cs.DC↗

arXiv2017

Transformation of Python Applications into Function-as-a-Service Deployments

New cloud programming and deployment models pose challenges to software application engineers who are looking, often in vain, for tools to automate any necessary code adaptation and transformation. Function-as-a-Service interfaces are particular non-trivial targets when considering that most cloud applications are implemented in non-functional languages. Among the most widely used of these languages is Python. This starting position calls for an automated approach to transform monolithic Python code into modular FaaS units by partially automated decomposition. Hence, this paper introduces and evaluates Lambada, a Python module to dynamically decompose, convert and deploy unmodified Python code into AWS Lambda functions. Beyond the tooling in the form of a measured open source prototype implementation, the paper contributes a description of the algorithms and code rewriting rules as blueprints for transformations of other scripting languages.

cs.DC↗

arXiv2017

Snafu: Function-as-a-Service (FaaS) Runtime Design and Implementation

Snafu, or Snake Functions, is a modular system to host, execute and manage language-level functions offered as stateless (micro-)services to diverse external triggers. The system interfaces resemble those of commercial FaaS providers but its implementation provides distinct features which make it overall useful to research on FaaS and prototyping of FaaS-based applications. This paper argues about the system motivation in the presence of already existing alternatives, its design and architecture, the open source implementation and collected metrics which characterise the system.

cs.DC↗

arXiv2017

Java Code Analysis and Transformation into AWS Lambda Functions

Software developers are faced with the issue of either adapting their programming model to the execution model (e.g. cloud platforms) or finding appropriate tools to adapt the model and code automatically. A recent execution model which would benefit from automated enablement is Function-as-a-Service. Automating this process requires a pipeline which includes steps for code analysis, transformation and deployment. In this paper, we outline the design and runtime characteristics of Podilizer, a tool which implements the pipeline specifically for Java source code as input and AWS Lambda as output. We contribute technical and economic metrics about this concrete 'FaaSification' process by observing the behaviour of Podilizer with two representative Java software projects.

cs.DC↗

arXiv2017

Exploiting the Cloud Control Plane for Fun and Profit

Cloud providers typically charge for their services. There are diverse pricing models which often follow a pay-per-use paradigm. The consumers' payments are expected to cover all cost which incurs to the provider for processing, storage, bandwidth, data centre operation and engineering efforts, among others. In contrast, the consumer management interfaces are free of charge as they are expected to cause only a minority of the load compared to the actual computing services. With new service models and more complex and powerful management abilities, it is time to rethink this decision. The paper shows how to exploit the control plane of AWS Lambda to implement stateful services practically for free and under some circumstances even guaranteed for free which if widely deployed would cause a monetary loss for the provider. It also elaborates on the consistency model for AWS Lambda.

cs.DC↗

arXiv2013

Operating the Cloud from Inside Out

Virtual machine images and instances (VMs) in cloud computing centres are typically designed as isolation containers for applications, databases and networking functions. In order to build complex distributed applications, multiple virtual machines must be connected, orchestrated and combined with platform and infrastructure services from the hosting environment. There are several reasons why sometimes it is beneficial to introduce a new layer, Cloud-in-a-VM, which acts as a portable management interface to a cluster of VMs. We reason about the benefits and present our Cloud-in-a-VM implementation called Nested Cloud which allows consumers to become light-weight cloud operators on demand and reap multiple advantages, including fully utilised resource allocations. The practical usefulness and the performance of the intermediate cloud stack VM are evaluated in a marketplace scenario.

cs.DC↗

arXiv2012

π-Control: A Personal Cloud Control Centre

Consumption of online services and cloud computing offerings is on the rise, largely due to compelling advantages over traditional local applications. From a user perspective, these include zero-maintenance of software, the always-on nature of such services, mashups of different applications and the networking effect with other users. Associated disadvantages are known, but effective means and tools to limit their effect are not yet well-established and not yet generally available to service users. We propose (1) a user-centric model of cloud elements beyond the conventional aaS layers, including activities across trust zones, and (2) a personal control console for all individual and collaborative user activities in the cloud.

cs.DC↗