Perumal is a Senior Researcher within Intel’s Labs Europe (ILE). He has been with Intel for 15 years, serving variety of engineering, operational and research roles.
Intel Corporation is the world’s largest semiconductor manufacturer, and develops advanced integrated digital technology platforms for the computing and communications industries. More than 6,000 researchers and scientists are engaged in our research and development program, including over 1500 employees within Intel Labs Europe.
The Ireland-based labs carry out their activities for Intel Research and Innovation Ireland Limited. These labs include the Cloud Services Lab. The Intel Cloud Services Lab (CSL) is focused on dependability and platform/service differentiation in cloud computing and virtualized and SDN-enabled networks; and on IT management framework processes for technology adoption. Specific research topics include instrumentation and manageability of cloud and network infrastructures, open interface specifications, service level agreements, context awareness and digital preservation. The lab’s research output includes usage models, prototypes, technologies and open standards which enable dependable cloud environments and support automatic discovery and selection of optimal platform services; as well as a process-based ￼framework for mapping IT improvement efforts. The Cloud Services Lab has helped advanced the state-of-the-art in research through numerous FP5, FP6 and FP7 engagements including T- NOVA, NETIDE, MOBILE CLOUD NETWORKING, IOLANES, FI-WARE, SLA@SOI, PERSIST, SOCIETIES, N4C, iSURF, DBE, as well as various joint initiatives with industry (SAP CoLab, BT CoLab) and academia (Innovation Value Institute).
RELEVANCE TO THE PROPOSAL
The Cloud Services Lab (CSL) as part of Intel Labs is chartered with both exploratory and directed research as well as near term path finding. The objectives on CSL’s research agenda are:
Cloud Service Dependability focused primarily on cloud platform interoperability, as well as exposing underlying infrastructure features (matched with automated discovery) that enable resilience.
Differentiate IA through Cloud Platform Services which make Intel platform innovations easily discoverable and consumable in the Cloud.
Drive Cloud Interoperability and dependability; through open extensible standards, protocols and capabilities.
CSL’s portfolio of projects comprises of applied R&D type projects that are grounded in enterprise IT current and strategic issues. The Cloud Service Labs has established a strong technical expertise in the cloud computing related projects through cloud related project in FP5 through to FP7. Output from the projects is used to inform architectural and engineering adoption and migration decisions within Intel’s enterprise environment.
TASKS ASSIGNED AND ROLE IN THE PROJECT
Intel will act as a contributor to WP2, WP3, WP4, WP6 and WP8. Intel’s key role in CloudLightning is the development of instrumentation that will be used to characterize the heterogeneous environment. Intel will provide expertise on MIC architectures and how it can be successfully integrated into cloud environments. Intel’s specific tasks are as follows:
• In WP2 Intel will work with UCC, NTNU, IeAT, Certh, and Maxeler to implement
representative use cases on the CloudLightning testbed. These use cases will test the validity and the performance of the self-organisation and self-optimisation management platform. Intel will use the instrumentation developed in WP6 to evaluate the performance of the use cases in the CloudLightning testbed and to determine how well the workloads perform, to identify bottlenecks and potential opportunities for configuration refinement.
• In WP3 Intel will contribute to architecture design of CloudLightning
• In WP4 Intel will develop methods to capture relevant data sets from the heterogeneous infrastructure environment (Xeon Phi, GPUs and FPGAs) which can be used to provide operational metric. These metrics will include both standard resource consumption metrics (resource utilisation, power consumption) and derive or hybrid metrics which combine multiple singular metrics to provide operational insights into how the platform is behaving e.g. job status, task queues, for a given workload.
• In WP6 Intel will develop instrumentation that will be used to characterize each co- processor type in order to capture its specific behavioral aspects. Secondly Intel will work with NTNU, Maxler and UCC to identify the support infrastructure required for each co-processor type. Finally, Intel will work with NTNU, Maxler and UCC to create a set of plugins that interfaces the heterogeneous co-processor environment with the self-organization and self-optimization management platform.
• In WP 8 Intel will with all the partners contribute to the development of an exploitation plan. Intel will also contribute the dissemination plan through engagement in publications, contributing content to the project website, participation in international conferences etc.