Cloud computing infrastructure and storage virtualization

Storage devices offer higher availability and lower costs in addition to better performance than switches or servers.

Cisco announced the "California State project", so that it directly into the server market, and the IT field on the switch-centric virtualization method debate has intensified. The project can be counted as a stroke, and if it succeeds, Cisco will erode most of the server market.

At the same time, the third area that may be more lucrative for architecture control points, that is, storage, has been overlooked as a server virtualization tool.

With high capacity requirements for the application environment and I/O intensive workloads, the storage-center approach delivers higher performance, lower costs, and higher availability than switches or server alternatives.

Traditional storage arrays are an integrated server virtualization technology because there is not enough computing available. Most SAN and NAS designs rely on two RAID controllers, which are typically proprietary CPUs with limited processing power. High-end Systems exchange High-performance X86 processors, but these controllers completely consume the need for raid protection and storage virtualization up to hundreds of drives. In addition, each controller must prepare a series of partners to prevent its controller from failing to run.

New extended storage systems will change market competition

By contrast, the new extended storage system overflows x86 resources. In these extended systems, the x86 power form storage device hardware storage area is expected to make management as a single capacity and performance. Specialized software runs the capacity and performance of each device in a parallel rollup, protecting data from any hardware failure.

Each storage device can have two X86 processors in the extended system, with a total ratio of one X86 processor to six drives. With this capability, the virtual server is running on existing X86 processors, allowing standard server applications to share hardware and shared storage.

This is the first time a user can consolidate a server application into a storage platform. As we all know, the benefit of server consolidation is to reduce hardware procurement costs, improve availability of virtual resources, and reduce management costs.

Reduce procurement costs

By running server virtualization on each device layer that extends outward, server applications can run on each device, and each application has access to the underlying array or appliance of the entire shared capacity and bandwidth. The server instance can start, stop, and manage any remote server.

By eliminating the hardware of the external physical server, users directly reduce the cost of the physical server along with the server power, cooling, maintenance, and rack space. The benefits of "green" are attractive, as power consumption is reduced by as much as 48% and can be achieved by separate servers and storage.

Higher availability

An externally extended storage architecture fails to allocate an equal number of devices with the inherent data device protection. A device failure is considered a simple raid activity, and data writes and reads are still unaffected.

Integrated virtualization server, if the device fails, the image device running on the virtual machine can automatically restart another available device array.

Due to the basic array protected by the virtual machine boot image, no user intervention occurs when restarting the application and re-establish the server virtual LAN connection and virtual MAC address. This can be done in minutes without the need for complex server clustering software, dedicated standby hardware, or specialized equipment interconnection.

Easy to manage

Virtualization simplifies general management by replacing physical devices with logical entities. Logical entities are expanded more easily, workloads can be automatically balanced, and system attributes are changed, because the dynamic management of virtualization is linked to a workload that requires material resources.

A common platform that provides both server and shared storage resources for physical devices, connecting standard Ethernet networks is the ultimate simplification and commercialization of hardware resources that has been predicted for many years. This approach provides a linear extension of computational and storage resources to prevent physical failure, as well as investment protection programs that are ideal for future private cloud computing platforms, where resources must be delivered and managed without fear of physical limitations.

Conclusion

Because virtualization offers such great benefits, customers should carefully examine the main workload environment and choose the best virtualization platform to meet their requirements. Rich I/O intensive work tasks, the new extended architecture based on the storage-centric platform can be further planned to consolidate the data center cloud computing infrastructure.