The expansion of smart devices, social networks and other applications in full diversification which now generate terabytes of data led the data center architectures to undergo a profound transformation. Existing infrastructure generally consist of software and hardware layers from multiple vendors, have become increasingly limited and ineffective. Each product requires its own user interface and training while being overloaded by its own components of computing and storage.
Virtualization infrastructure to hyperconvergées
The response of the IT industry this growing deficit was scalable. Virtualization, in which virtual machines (VMs – virtual machine) can be moved around real server hardware has become ubiquitous, with over 80% of virtualized workloads in 2016, according Gartner1. Consolidating multiple VMs on a single physical server reduces hardware costs and footprint, and power and cooling requirements. Other benefits include better application availability, reduced downtime, increased productivity and speed, and a business continuity and disaster recovery optimized.
In addition to its many benefits, virtualization also presents challenges, especially when it involves the sharing of resources for storage or networking. The emergence of converged infrastructure (CI) enables this challenge to some extent by combining the storage and server components into a single device so as to not require a storage area network (SAN – Storage area network). The infrastructure virtualization complements converged, by virtualizing the storage layer so that it runs on the virtualization platform. It proved essential for IT by enabling simplified management, advanced automation, rapid application deployment and organizational agility. However, it is somewhat limited. A converged infrastructure is a fixed system with a defined level of resources allocated for the calculation, storage and networking. This configuration may prove to be less flexible than what some companies require; it simply migrates to the storage network virtualization platform instead of the challenge of the underlying data management. The hyperconvergence is the next step towards achieving the ultimate goal of a software-based data center (SDDC – software defined data center).
A hyperconvergée infrastructure solution (HCIS – hyperconverged infrastructure solution) includes devices stacks where computing resources, storage, network and virtualization and other technologies are fully integrated. It allows IT professionals to reduce the complexity of their IT environments, reduce risk and avoid having to evaluate every hardware and software. In addition, its small footprint reduces power and cooling costs.
energy management complete data management
These technologies (virtualization, converged infrastructure and now HCIS devices) significantly improve the performance of data centers and are viable solutions to meet the current requirements in performance and flexibility. However, this potential will quickly be compromised if their implementation is not accompanied by a coherent energy management strategy and also advanced. Virtualized servers host multiple virtual machines; they therefore operate at a capacity of 70 to 80% compared to that of a non-virtualized machine (10 to 15%). Each unit also requires more energy, up to 40 kW. In addition, virtualization means that applications (and their energy requirements) can be moved from one server to another quickly and impromptu, or even between data centers.
Therefore, a successful virtualization requires providing the right energy in the right place at the right time. It allows not only to avoid costly energy overload, but also to ensure the quality and availability of energy. Indeed, a recent report suggests TechRadar that the real cost of the unavailability of a data center could be at least $ 10,000 per minute2.
Optimizing energy management is essential in high-availability environments where unplanned outages or line quality irregularities may have a negative effect on computer applications. Businesses require intelligent and reliable energy management strategy to exploit the advantages of modern computer architectures, improve efficiency and reduce the risk of business interruption.
hardware and energy management software
Eaton works closely with the leaders of virtualization technologies, CI and HCIS, such as VMware, Citrix, Microsoft, NetApp, EMC, VCE, Nutanix, SimpliVity and Cisco, to provide energy management solutions validated in laboratory and compatible with modern virtualized architectures. Eaton also provides reference patterns for a wide range of converged solutions and hyperconvergées available on the market.
The Intelligent Power Manager (IPM) Eaton, for example, can be integrated into the main virtual machine management systems, including VMware® vCenter ™, Microsoft SCVMM ™ and Citrix XenCenter ™, providing IT managers with an environment of unified control and monitoring including data infrastructure and energy.
In case of power failure or weather phenomenon, the affected computer equipment can be stopped cleanly and automatically to save the work in progress and preserve data integrity. Virtual machines can be migrated to other locations or tipping sites. The IPM software has a feature intelligent load shedding, which can suspend non-critical virtual machines, thus improving system availability while increasing battery life and reducing the load on the generator.
Among its many functions, the level of consumption demand keeps running critical workloads in the event of power failure by limiting the power consumption of the server. This can help to lengthen the execution time up to 200% with the same number of inverter battery modules associated with an integrated load shedding function. Moreover, the inverters and 5PX 9PX ready for virtualization and Energy Star Eaton, with a capacity extended battery, provide a versatile backup power, efficient and reliable, while the racks management options and cables maintain the integrity of the equipment, improve air circulation and reduce cooling costs.
In addition, for energy distribution units (ePDU Eaton G3) have evolved into true modules intelligent power distribution, indicating energy consumption and efficiency with high accuracy and taking decisions on the capacity planning and reducing consumption. An advanced LCD screen, a hot-swap counter and operating at high temperatures (60 ° C) are some of its other functions.
Accurate monitoring of energy allows load balancing, identification of unwanted energy losses and locations with an energy capacity of relief. In addition, the energy management solution Eaton can determine the ecological threats such as peak temperatures or humidity, receive security alerts and notifications adapt. Combined with intelligent distribution unit of energy (ePDU G3), the environmental monitoring sensor (EMP – Environmental Monitoring Probe) Eaton can detect and explain these conditions to enable automatic migration of virtual machines to cloud or backup sites.
Today hyperconvergée advanced technology to meet the essential and demanding computing challenges quickly and economically. However, the data management strategy computer rooms must be accompanied by an equal quality of energy management strategy to meet the expectations of their users reliably, even under adversity or disaster situations . For this, Eaton has developed energy management components reducing the time and complexity of deployment, reducing risk, lowering total cost of ownership and ensuring business continuity and data integrity.