direkt zum Inhalt springen

direkt zum Hauptnavigationsmenü

Sie sind hier

TU Berlin

Inhalt des Dokuments

Highly-optimized Cloud Infrastructures

Lupe

A solid technical foundation on Cloud Computing was developed over the last years ranging from resource allocation on the Infrastructure as a Service (IaaS) layer, job and application execution on the Platform as a Service (PaaS) layer, to software and service usage on the Software as a Service (SaaS) layer. This foundation allows the access of resources according to the pay-as-you-go principle. At this, the resources are located in datacenters provided by companies all around the world and users can access them remotely without the need for any upfront capital expenses, long-term commitment, or significant provisioning delay. However, this cloud concept brought many new challenges with respect to Service Level Agreements, Software-defined networking, Energy awareness, Scalability, Availability, and Reliability.

 

Infrastructures

Our group manages more than three computing clusters to conduct state-of-the-art research for highly-optimized cloud infrastructures. The biggest cluster consists of 200 nodes and is used for cloud computing research, focusing on scalability of algorithms and system management procedures. Two currently running projects make use of a smaller cluster compound of 16 cluster nodes. We prosecute energy-meters to capture the individual power consumption of each cluster node. In the ASCETiC project, we use this cluster to improve energy-efficiency of cloud services. It is also available for young start-ups of the TU Berlin in terms of the CCI project. We use OpenStack to manage this cluster's infrastructure. A third cluster consists of nearly two terabytes of RAM and was primarily used in the EC-funded EFRE project. It consists of ten cores with two six-core CPUs in each.

Service Level Agreements

Service Level Agreements (SLA) are established and managed between Cloud Service Provider and Cloud Service Consumer and are an instrument to describe a service and its quality. This topic is of crucial importance for companies whose success is depending on the advertised Quality of Service (QoS) that the service provider delivers. Thus, an innovative way have to be found for SLA management and monitoring of guarantee fulfillment comprising agreement template definition, agreement negotiation, agreement creation and of course the validation and evaluation of agreements. The work group of Complex and Distributed Systems (CIT) strongly conduct research in this domain which is exposed by the successful projects listed below on the one hand and is reflected in the high number of research papers on the other hand. Furthermore, we delivered not only the Web Service Agreement Negotiation (WS-Agreement Negotiation) specification in cooperation with the Open Grid Forum (OGF), but still working also on the Web Service Agreement Framework for Java (WSAG4J).

Research projects

Current research topics include:

  • Network-as-a-Service
  • Service Level Agreements
  • Software-defined networking
  • Energy awareness
  • Virtualization
  • Infrastructure-as-a-Service
  • Scalability, Availability and Reliability
  • Quality of service
  • Mobile networking and services
  • Hardware-as-a-Service

ASCETiC

Lupe

ASCETiC focuses on providing novel methods and tools to support software developers aiming to optimise energy efficiency and minimise the carbon footprint resulting from designing, developing, deploying, and running software in Clouds. Learn more...

AssessGrid

Lupe

The project AssessGrid adresses the risk awareness and consideration in SLA negotation, self-organising fault-tolerant actions, and capacity planning in the context of Grid computing. Learn more...

BCI - Berlin Cloud-based Infrastructures

Lupe

The project BCI (Berlin Cloud-based Infrastructures) is supported by the European Regional Development Fund. Its focus is the realization of a cloud-based infrastructure solution for start-up businesses. Learn more...

Capital Cloud Infrastructures (CCI)

The Capital Cloud Infrastructure offers established start-ups of the Berlin Capital Cloud the opportunity to switch to a commercial cloud computing provider with low adjustment efforts and service downtime. Therefore the focus of the project is to design a mechanism for an automatic migration of data and virtual machines between these cloud environments. Learn more...

Condel 2

The Condel 2 projects is working on the improvement of optimistic algorithms to make distributed, consistant decisions in huge and error-prone networks with varying participants. Learn more...

HPC4U

Lupe

The goal of the project HPC4U is to provide a predictable and reliable middleware for clusters built of Components-off-the-Shelf. A major research focus is put on checkpointing and job migration mechanisms, which allow for an SLA-compliant job execution even in the event of hardware failures. Learn more...

MO-BIZZ

Lupe

MO-BIZZ focuses on an ecosystem for new mobile, cloud-based services which are highly flexible, resource-optimized, and SLA-aware provided by autonomous organizations. This system encompasses and integrates three major domains: Network and Computing Infrastructures, Mobile Networks and Cloud Services. Furthermore, a business model for such rapid service development environment allows businesses to expose their key functionalities to their customers. Learn more...

OFELIA

Lupe

The goal of the OFELIA project is to create a unique experimental facility based on the OpenFlow technology that allows researchers to not only experiment "on" a test network but to control and extend the network itself precisely and dynamically. Learn more...

SLA4D-Grid

Lupe

The project SLA4D-Grid is working on the design and realization of a Service Level Agreement layer for Germany's national Grid infrastructure D-Grid. The new layer offers the different actors of the D-Grid the service usage under given guarantees, quality-of-service requirements and pre-defined business conditions. Learn more...

TVPI - T-Labs Virtual Project Infrastructure

The T-Labs Virtual Project Infrastructure (TVPI) project is a cooperation between the CIT and the Telekom Laboratories (T-Labs) at TU-Berlin. It aims at implementing and running a cloud-based IT infrastructure that supports upcoming software development projects inside T-Labs. Learn more...

Zusatzinformationen / Extras

Quick Access:

Schnellnavigation zur Seite über Nummerneingabe

Auxiliary Functions

Contact

Odej Kao
+49 (30) 314-24230
Room E-N 158

Contact

Alexander Stanik
+49 (30) 314-22616
Room E-N 174