Alexander Lenk

What's inside the Cloud? An architectural map of the Cloud landscape

We propose an integrated Cloud computing stack architecture to serve as a reference point for future mash-ups and comparative studies. We also show how the existing Cloud landscape maps into this architecture and identify an infrastructure gap that we plan to address in future work.

Cloud service engineering

Building on compute and storage virtualization, Cloud Computing provides scalable, network-centric, abstracted IT infrastructure, platforms, and applications as on-demand services that are billed by consumption. Cloud Service Engineering is the application of a systematic approach to leverage Cloud Computing in the context of the Internet in its combined role as a platform for technical, economic, organizational and social networks. This tutorial introduces concepts and technology of Cloud Computing and Cloud Service Engineering, providing an overview of state-of-the-art in research and practice. We show how to set up a private Cloud that delivers Infrastructure-as-a-Service (IaaS). Eucalyptus and OpenNebula are popular open source software frameworks for creating on-premise Clouds. Promises, challenges and solutions for integrating services of a private Cloud with public Cloud services such as Amazon EC2 and SQS are discussed. We show how the best of both worlds - private and public Clouds - can be combined to build scalable and secure systems.

Cloud Application Portability with TOSCA, Chef and Openstack

As the sector of SaaS is being evolved, the need of portability and effective orchestration of Cloud-enabled applications over virtualised infrastructures is more apparent. In this paper we present a proof of concept (PoC) project together with industry partners which deals with the management of portable Cloud applications. Experimentation with a multi-tier Cloud application use case gives insights into the capabilities of the TOSCA specification (version 1). To realize support for a Cloud environment based on OpenStack and Opscode Chef, an environment that allows the execution and management of TOSCA service topologies has been developed. For the evaluation of the implemented system, we orchestrate the un-/deployment of the multi-tier Cloud application use case and in the end we discuss the results and experiences of this exercise.

Requirements for an IaaS deployment language in federated Clouds

The advent of advanced virtualized IT infrastructures that can be provisioned as on-demand services, known as Infrastructure-as-a-Service (IaaS) Cloud Computing, has created new research challenges and opportunities. The capability to rapidly allocate and deallocate seemingly infinite amounts of system resources is a defining characteristic of this technological trend. Operating multi-tier applications on a continuously changing environment is one of the big challenges in IaaS Cloud Computing. This challenge is even more daring, if applications are not running at one specific Cloud site but on multiple different sites of various providers. This work is motivated by the need for new deployment description approaches that target application run-time aspects in federated Clouds. We propose six key requirements for IaaS deployment description languages that facilitates continuous application deployment on a permanently changing infrastructure across multiple Cloud sites: software deployment on dynamic virtual machine resource pools, continuous system supervision and change management, a generic model for federated Clouds, automated software configuration management, multi-tier dependency management, and use of a machine-readable language.

Cloud Standby: Disaster Recovery of Distributed Systems in the Cloud

Disaster recovery planning and securing business processes against outtakes have been essential parts of running a company for decades. As IT systems became more important, and especially since more and more revenue is generated over the Internet, securing the IT systems that support the business processes against outages is essential. Using fully operational standby sites with periodically updated standby systems is a well-known approach to prepare against disasters. Setting up and maintaining a second datacenter is, however, expensive. In this work, we present Cloud Standby, a warm standby approach for setting up and updating a standby system in the Cloud. We describe the architecture of Cloud Standby and its methods for deploying and updating the standby system. We show that by using Cloud Standby the recovery time and long-term costs of disaster recovery can significantly be reduced.

Modeling Quality Attributes of Cloud-Standby-Systems

Contingency plans for disaster preparedness and concepts for resuming regular operation as quickly as possible have been an integral part of running a company for a long time. Today, large portions of revenue generation are taking place over the Internet and it has to be ensured that the respective resources and processes are secured against disasters, too. Cloud-Standby-Systems are a way for replicating an IT infrastructure to the Cloud. In this work, a Markov-based model is presented that can be used to analyze and configure such systems on a long term basis. It is shown that by using a Cloud-Standby-System the availability can be increased, how configuration parameters like the replication interval can be optimized, and that the model can be used for supporting the decision whether the infrastructure should be replicated or not.

TIOSA: Testing VM Interoperability at an OS and Application Level -- A Hypervisor Testing Method and Interoperability Survey

Virtualization is the foundation of modern, cloud-based applications. The existence of virtual machines (VM) that host the components of such applications enables their portability and scalability. VMs are used in cloud infrastructures, and with dynamic operational requirements there is a need to move VMs within and across different clouds. The successful migration of VMs from one cloud to another should not always be considered as given. The goal of this paper is to devise and implement methods and conduct functional tests towards evaluating interoperability in cloud environments. We suggest a methodology for assessing the interoperability across different systems and we conduct a survey with a series of hypervisors and operating systems.

Cloud Standby Deployment: A Model-Driven Deployment Method for Disaster Recovery in the Cloud

Disaster recovery planning and securing business processes against outtakes have been essential parts of running a company for decades. As IT systems have become more important, and especially since more and more revenue is generated over the Internet, securing the IT systems that support the business processes against outages is essential. For cost reasons, we proposed Cloud Computing for disaster recovery in prior work. In order to update and deploy a standby system, however, a deployment method that allows a provider independent, automated deployment of distributed systems is necessary. In this work we present such a deployment method. We show that by using Cloud Standby Deployment, the recovery time can be reduced significantly, and the distributed system can be deployed provider independently.

Vertical Scaling Capability of OpenStack

With the emergence of cloud computing, resources can be dynamically scaled. A common scaling approach is the addition and removal of virtual machines, known as horizontal scaling. Horizontal scaling can take several minutes but erratic and sudden changes in demand take place within seconds. Therefore, vertical scaling has been introduced, changing the resources of an existing virtual machine during run time and within one second or less. At the same time, more and more private clouds and cloud providers apply the open-source platform OpenStack. Hence, this paper evaluates the vertical scaling capability of OpenStack. For this purpose, we examine whether and to what extent common guest operating systems, popular hypervisors, and OpenStack itself support vertical scaling. Altogether, the considered operating systems and hypervisors support vertical scaling of almost all considered resources while OpenStack does not support vertical scaling at all. Based on our findings, we finally suggest steps to improve OpenStack.

An Open Framework to Support the Development of Commercial Cloud Offerings based on Pre-Existing Applications

Web-based service delivery and billing by consumption are two defining properties of Cloud Computing. They leverage novel business models and sales channels for yet to develop as well as pre-existing applications. Many pre-existing applications are already capable of running in distributed environments, but still do not meet the requirements to run as a Cloud offering. For example, applications built to run in cluster environments are designed for distribution and massive scalability and from an architectural point of view qualify to be suitable for Cloud environments as well. However, a limitation of cluster-based applications with respect to Cloud adoption is that many of them initially do not support necessary Cloud service features such as payment services or multi-tenancy. To close this gap, we propose an open framework that helps to adapt pre-existing applications into commercial Cloud offerings. The framework facilitates the process to extend pre-existing applications with the respective required web service interfaces, which ultimately allows them to be consumed as Cloud offerings in a pay-as-you-go manner. The framework approach is illustrated by the transformation of the two well-known cluster applications Hadoop MapReduce and MySQL into full-blown Cloud offerings.