Intel Cloud Builders Reference Implementation
A Reference Implementation is developed against an Intel Cloud Builders reference architecture. It describes in detail an implementation in a real world customer environment and further demonstrates the intent of reference architecture.
Implementing mission-critical business software such as SAP's Enterprise Resource Planning solutions on virtualized and cloud computing platforms requires a good understanding of infrastructure performance in order to be able to meet critical business needs. Capgemini, a leader in consulting, outsourcing, and professional services demonstrates in this paper how a good match of processor and infrastructure capabilities is able to flexibly handle the capacities required for different sizes of SAP environments. Through the combination of the reliability, availability, and serviceability (RAS) features of the Xeon E7 platform with storage and virtualization features aimed at high availability and disaster recovery, this virtualization platform represents a good foundation for highly resilient cloud services for business.
GreenField Health provides consulting team services to a wide range of customers focused on new practice start-up, healthcare quality, performance improvement, information technology, system design, and cost management, and was seeking to better utilize power in existing data centers. The company decided to do a proof of concept of an out-of-band with
policy engine available with the Intel® Xeon® processor 5500 and 5600 series and Intel® Node Manager.
GreenField expressed a desire to not only monitor power, but also to manage power consumption. Recognizing current trends in power costs and utilization are increasing, GreenField knew that measuring power consumption levels and having the ability to cap power would provide the next level of control in cost management. For a trial of power capping, GreenField implemented an Intel Node Manager solution encompassing one server at a data center facility and running virtual machines with the patient care applications and assessed power efficiency in a new server configuration.
This reference implementation demonstrates application deployment using HP CloudSystem Enterprise, Intel® Xeon® Processor E5 Family, and Software-as-a-Service (SaaS) products. The purpose of this document is to describe how to set up and take advantage of the HP CloudSystem Enterprise features in order to deploy a two-tier Linux infrastructure on both virtual and physical servers.
With HP CloudSystem Enterprise, you can get the benefits of a full service deployment for both private and hybrid cloud environments. HP CloudSystem is the product of HP's experience in delivering industry-leading automation, application management, and Converged Infrastructure capabilities. It enables organizations to build and manage cloud services across private clouds, public clouds, and traditional IT. As part of HP's Converged Cloud architecture, clients have a simplified, integrated architecture that is easier to manage and provides flexibility and portability between private, public, and managed clouds.
The audience for this Intel® Cloud Builders implementation guide includes enterprise IT and Line of Business organizations deploying private and/or hybrid cloud solutions to transform their internal data center infrastructure into a cloud environment with the goal of offering better services to their internal customers or users. Additional audiences include systems integrators looking to transform the cost structures and agility of their client offerings by providing cloud computing services and solutions that accelerate revenue potential by reducing time to market, and cloud service providers that offer public and hybrid services to consumers.
This reference implementation was compiled to assist IT professionals considering a power management solution in order to achieve data center power efficiency, reduce power costs, and to improve continuity of service in the event of a disaster.
The power capping function of Intel® Node Manager, available in the newer Intel® Xeon® processor E5 family, can be used to impose power consumption targets on servers. Intel® Data Center Manager implements the orchestration of power policies across large numbers of servers. These policies can implement a data center-wide peak shaving strategy to ride out peak consumption limits during the high demand summer season.
By combining Intel Data Center Manager with Intel Node Manager, there was a significant impact on power savings, and the technology can be implemented to manage data center power effectively. The promise of this technology is that with an SLA measurement capability, it would be possible to reduce energy consumption by about 10 percent during the high demand summer season.
RACSA (Radiográfica Costarricense, S.A.) deployed a series of servers at the RACSA-Zurquí hosting facility in Costa Rica for an initial proof of concept (POC). The goal was to provide RACSA a framework to evaluate the applicability of Intel® Intelligent Power Node Manager (Intel Node Manager) on Dell* C 2100 series servers in the current RACSA data center environment from both a technology and an IT process suitability perspective. The pilot allowed RACSA to assess power management technologies under a controlled environment, using the embedded platform instrumentation capabilities of Node Manager enabled servers in combination with the appropriate IT processes. Intel Node Manager allows real time readouts of server power consumption, and allowed RACSA to analyze performance and actual power consumption and assess energy consumption in energy efficient platforms.
RACSA saw immediate benefits from the instrumentation embedded in Intel Node Manager-enabled Dell C-series servers. The availability of real time and historical power consumption figures enabled the system administrator to provision servers to actual power consumption without impacting performance and within the boundaries of the limited available power in a containerized environment.