Blade Server--best choice for data center

In the era of cloud computing, blades are the base server for aggregation systems and will become part of the private cloud, and critical business systems need to adapt to the needs of the cloud-era users. Superdome2 as the representative of HP Integrity Server began the innovation process, blade expansion of the architecture of the adoption, standardization and modular choice of structure, these allow the integrity kinetic energy Blade Server has an extraordinary vitality and perfect expression, easy to deal with customer critical requirements, Let customers more resources for business innovation.

In fact, the three main advantages of blade servers coincide with the needs of cloud Computing's basic platform architecture-high-density virtualization applications, highly flexible scalable dynamic architectures, and higher energy efficiency. But in key business areas, Blade Server application is still in the initial stage, on the one hand, because the early UNIX blade server in terms of scalability, reliability is not good enough, in the hardware itself is congenitally deficient; On the other hand, it is the user to the key business areas of the blade also exists misunderstanding, There are doubts about the blade of a business-critical server.

Blade server breaks critical business hardware bottlenecks

UNIX is still the main platform to support the core applications of large enterprise users, but before HP releases the integrity Kinetic Blade server based on Itanium 9300, the industry lacks blades that enable vertical scaling and the I/O ports and bandwidth needed to deliver core applications. For example, HP Integrity bl870c, the highest-end blade server in the Unix world, can only support 4-way 8 nuclear processing power, apparently not enough to handle large applications. This leads to a number of user misconceptions-because the user needs a high-performance single system, is a "high-performance node", the key Business Server "natural" should be rack-type.

This recognition has become obsolete after HP launched the Integrity BL8X0CI2 series Blade Server based on the Itanium 9300 processor--A integrity Bl890ci2 blade server that can hold 2 8 CPU slots in a C7000 chassis that occupies 10U space. The maximum number of memory per Bl890ci2 can also reach 384GB. It can be said that Integrity890ci2 has reached and even surpassed HP's previously Advanced Server rx7640 and rx8640 in terms of CPU volume and memory capacity, and has been able to seamlessly connect HP's most upscale integrity Superdome or Superdome2 server.

The blade server has always been a weak I/O scalability, the next generation of I2 series blades are also well resolved-the lowest end of the BL860CI2 standard up to 4 10GB Gigabit Ethernet ports, And there are 3 pci-e2.0 mezzanine card expansion capabilities, can be used to increase the dual-Port FC card, Infinite band and other ports, while the highest BL890CI2 standard 16 10GB Gigabit Ethernet port, 12 pci-e2.0 expansion slots. Each 10Gb port can also be converted to 4 1Gb Ethernet ports via the Hpc-class Flex-10 module. Therefore, theoretically bl890ci2 can support up to 64 Gigabit Ethernet ports and 12 expansion slots without any extensions, which is clearly beyond the configuration required by most users in the high-end servers. Clearly, HP's I/O design is aimed at a high intensity core application and integration infrastructure considerations.

As can be seen from the above description, simply from the hardware configuration, HP Blade server is enough to meet the user's key application needs. So is this progress in hardware configuration at the expense of RAS? Obviously not.

Make sure the RAS blades are more flexible

RAS is an important reason why UNIX servers can be far ahead of x86 servers in key applications. HP blade structure of the key Business Server porting the original rack-type minicomputer all of the high reliability, high availability and redundancy design, and has a higher scalability, completely in accordance with the user's key business to the highest requirements of minicomputer design. For example, each node of the new integrity Kinetic Blade Server has 2 CPU slots and has its own memory, I/O channel.

Not only that, HP also provides more than 100 improvements in the new generation of integrity Kinetic servers, with its built-in business continuity, stability, and usability capabilities to keep business uninterrupted and performance up to 2 to 9 times times. For example, HP's dual-chip standby feature provides a 17 times-fold higher memory reliability than a single chip standby. The Superdome server, which also uses blade design, can not only ensure successful transaction tracking, but also reroute transactions by the second path, and supports I/O and CPU independent extensions. It should be noted that the power and fan prices are low, but this is the most vulnerable component in the server. HP's blade structure will separate the two parts, put in the knife box, and can be hot-swappable, which also greatly improve the overall system of sustainable operation.

More importantly, with the same chassis as the x86 server, this gives users greater flexibility in deploying the IT infrastructure. Based on the experience gained in engineering design over the past few years, HP has released key business integration infrastructure, merging different computing environments and communication devices, and opening up all it islands in the user's environment. Behind this new architecture, HP has not only achieved a unified blade architecture from x86 to Superdome, but has built an elastic structure based on more than 100 innovative technologies to cope with various types of workloads.

A user integrated SAPSD3 layer solution in a C7000 enclosure fully embodies the management benefits of a converged infrastructure architecture: within 1 C7000 chassis, 2 installed INTEGRITYBL870CI2 server-HP-UX clusters are deployed, one for quality assurance, Test server and failover data server, another as the primary database server; The remaining space of the C7000 can be deployed to use the ProLiant bl460c Blade Server installed with SuSE as the SAP application server. This fully embodies the benefits of modularity, multiple operating system flexibility, and a unified management model greatly reduces management difficulty.

It should be noted that Hewlett-Packard in the latest integrity Bl8x0ci2 Blade Server also implemented a rack expansion: Users can be 2 sets of integrity Bl860ci2 simply connected together into a integritybl870ci2; The 4 integrity Bl860ci2 can also be turned into a 8-way integritybl890ci2 by this simple connection, which means that users not only maximize investment protection, but also have a great deal of flexibility in the use of infrastructure.

In fact, the integration infrastructure that HP advocates includes not only servers, data networks, but also integrated storage systems and SAN storage networks in the same blade architecture. At present, integrity Kinetic energy blade Server can not only with other x86 server and storage blades, all the current C-class virtualization, unified management features can support, including virtual Connect. It can be said that HP's key business strategy has entered the level of integration infrastructure, not a simple product, device support. For users, critical business integration infrastructure means that users can use a unified architecture to manage critical business applications, and users can choose different operating systems and different servers on the same hardware platform at their own SLA level, which means that users have a greater choice of application space, Also greatly reduces the cost of users in key business applications.

Talking about TCO Blade winning rack

When it comes to cost, the test data shows that the blade server can consume less than 1/3 power consumption compared to the equivalent performance of the rack, which has not been calculated to calculate the relative heat dissipation. Through the expansion of the network, HP Virtual connection technology can achieve a connection, online adaptive new connections, greatly enhance the network ductility and configuration flexibility, can increase the maximum bandwidth of 20 times times. This means that users only need to configure network settings once, and they can manage all network reconfiguration in a virtual way and reduce the number of cables required by 97%. It can be said that the blade server in energy-saving, cable management and other aspects of the advantages have been the industry's consensus.

On the new generation of integrity, because of the integration infrastructure architecture, the reuse of many x86 servers in the common accessories, users on UNIX servers in the total cost of IT infrastructure is much lower than the traditional rack-type UNIX server. At the same time, through the unified management method, the management cost also obtains the very big reduction. "Integrity blade servers are managed using Hpsystems Insight Manager (SIM) software," said Eric Castaneda, Ferguson Company's UNIX services Manager. Hpsim manages various types of servers, storage devices, and network solutions for different operating systems through a single point of control. HP's Integrity Blade server runs the Hp-ux11iv3 operating system, which can be split into separate parts to perform operations without weakening performance and redundancy. ”

According to HP's test data, in the same 32-core configuration to run the database application, the blade design of Integrity Superdome2 3-year total cost of ownership is lower than the previous generation of Superdome 50%: including daily maintenance costs reduced by 57%, Software licensing costs are reduced by 50% and the initial purchase cost of the hardware is substantially reduced.

The core application blade has been burdened

It is not news that customers are applying blade servers in their core applications. "To stay competitive, users need a flexible IT environment that improves management efficiency and security while reducing costs." Users across the Asia-Pacific region expect HP's technology and product innovations to improve their productivity. HP Blade systems based on Itanium 2 processors and HP-UX operating systems meet this business requirement for users with simpler and more flexible blade solutions. "said Kirk, the Hewlett-Packard Company.

Now, with HP's help, the complex application infrastructure will only take 4 hours to complete the installation and function, rather than the number of weeks that are currently required, and by increasing the disaster recovery rate by 80%, it takes a few minutes to complete the failover, which greatly reduces server recovery time and can change from a few hours to a "minute" level, and save 80% of server equipment costs. Harald Fischer, director of the man Nutzfahrzeugeit Center, said: "At present, we have a HP Integrity Superdome server based on the Intel itanium® architecture, and have found a lower cost, flexible blade server that can easily install, Supply and manage different operating systems. The HP Integrity Blade Server meets our needs, allowing us to install UNIX applications in the same chassis as existing Windows and Linux applications, fully integrating hp-ux11i virtualization and management capabilities to achieve seamless integration. ”

"Ferguson has fully recognized the obvious advantages of blade solutions," said Eric Castaneda, Ferguson's Unix services manager. By reducing the cable to simplify management, we save time and save money. HP Integrity Blade Server will allow us to deploy HP-UX applications in the same chassis as existing Windows and Linux applications. This seamless integration and anticipation will drastically reduce TCO, allowing us to consider using HP integrity blades in future device installations. ”

Zhang Honglei, director of the Enterprise Informatization Department of Shandong Telecom Branch, sees HP's converged infrastructure to meet the needs of telecoms companies in building cloud computing platforms. "HP's blades are not a simple overlay of the usual servers, but a collection of modular components that integrate multiple modules to intelligently share the infrastructure." "This design allows HP to have a certain degree of redundancy inside the blade chassis," Zhang Honglei said. At the same time, Blade Server features also make it very easy to set up redundancy between blade servers and set up some resources as back-up. In other words, under the framework of the blade server, it is very easy to implement a blade server to provide redundancy to multiple blades, which is more efficient and better than the traditional dual-machine hot standby method. ”