Pages

4 February 2015

CREATING RESPONSIVE, ENERGY EFFICIENT COMPUTING CLOUD

FEBRUARY 3, 2015
A recent symposium on software performance gave IT researchers the opportunity to discuss progress made on the EU-funded CACTOS project. Over 40 researchers and practitioners from across Europe attended the symposium in Stuttgart, Germany to share results and practical experience.

The key objective of CACTOS, due for completion in September 2016, is to enable data centre operators to deliver cloud-based applications – using current and future hardware – as energy efficiently as possible. The project, which will receive a total of EUR 3 450 000 in EU funding, plans to deliver a set of tools and methods to help data centre operators analyse hardware and software behaviour and overall infrastructure performance.

One of the tools highlighted during the symposium was CactoSim, which is designed to enable users to simulate optimisation models, predict the behaviour of applications on different resources and validate and improve models. By enabling users to model and adopt the best configuration for their needs, the tool should help data managers to lower costs and achieve greater energy efficiency.

Cloud computing is a little like an electricity grid in that it is fundamentally about sharing resources in order to achieve economies of scale. In cloud computing, different services — such as servers, storage and applications — are delivered to an organisation or individual’s computer or device through the Internet, like, for example email. Many people now rely on cloud computing for a majority of their applications.

The current availability of high-capacity networks, low-cost computers and storage devices as well as the widespread adoption of hardware virtualisation, service-oriented architecture, and autonomic and utility computing have led to a growth in cloud computing. In order to keep up with this rapid expansion and still operate efficiently and cost-effectively, data centres handling all this cloud computing data have had to make significant investments in energy efficient buildings, server racks and facility management technology.

Given the scale of their operations, it is critical that business decisions such as these are made with as much information as possible. This then is the key objective of the CACTOS project. All operations carried out in a cloud environment are being mapped, so that workloads can be attributed to the most appropriate resources. This means in effect that information can be sent to the best fitting data centre at a given time. In case of failure, the next best matching place will be automatically detected and the workload relocated.

Currently ongoing research concerns power consumption analysis of software systems, and the accurate sizing of data centres. The consortium also plans to collect data on the actual use of cloud infrastructure, in order to help prioritise resources.

The final CACTOS toolkit – which will include the CactoSim tool – will be validated against three distinct scenarios for business analytics, enterprise applications and technical computing use cases. The end result will be the creation of a cloud computing environment that is more energy efficient, intelligent and responsive to user needs.

No comments:

Post a Comment