wholesale jewelry gift boxes and shipping materials What are the differences between VDI's VDI requirements and virtualization

wholesale jewelry california VDI allows the servers deployed inside the data center to deliver complete desktop instances to many devices, including traditional PCs, thin clients, and even zero client devices. However, each VDI instance is processed and stored by the server. Even less instances may require a large number of computing resources and network access. VDI deployment must start with careful consideration of server performance and evaluate the needs of server hardware upgrade. This article explains some common server problems related to VDI hardware requirements. It is necessary to point out that there is no only VDI hardware demand list. The problem is not lack of support. VDI can almost run on any virtual server. However, the number of VDI instances that can be deployed on the server is limited by the server's available computing resources.
For example, a typical "white box" server for enterprise -level VDI deployment may include two 8 -core processors and at least 192GB of DDR3 memory. The storage used by VDI instances is likely to be concentrated SAN storage. However, in order to avoid storage and VDI traffic appear in the same local area network, SAN may use separate networks (such as FC or separate physical local area networks) or local storage loading and protecting VDI instances on local storage on VDI servers. This means that the VDI server may require 16 SAS 6Gbps high-performance hard disks with a speed of 10-15K (height may be 2U or 3U).
Preading servers can support more VDI instances, while the number of VDI instances supported by server with earlier or poor performance is small. The server configuration listed above may be able to support 80 to 130 instances. However, the accurate quantity of the VDI instance that the server can support depends on the size and complexity of the basic mirrors, the degree of personalization, the number of virtual applications, the user in the Internet, users in the network And the degree of activity of the application.
It looks like there are many examples, but a large -scale enterprise that adopts VDI may hire 1,000 employees or more -this means that at least 10 servers are required, and additional servers need to be used for at least 10 servers. Support the number of instances increase and fault switch. Enterprises with 5,000 users will need about 50 such physical servers. The number of physical servers increases, and the cost of the VDI platform will increase accordingly. VDI completes all processing tasks in a server, and only uses terminal devices as an I/O platform (such as video, mouse, keyboard). Therefore, all desktops and visual rendering work are completed in the processor of the host server. The generated images are transferred to the terminal device through the local area network. The basic Windows desktop session and other elements usually have no problems, but it is possible to encounter problems when performing advanced graphic tasks (such as streaming media or 3-D graphics).
The question is precisely hardware support. The server often omit GPU, because traditional server -side tasks such as file server or activity directory server do not use graphics. However, when the graphics instruction (such as the SSE3 instruction) needs to be processed, the GPU cannot be uninstalled with the GPU -only the CPU uses unparalleled software simulation to get these instructions. The result is that the performance is significantly reduced, and all VDI instances related to the affected CPU core will be affected. Because VDI uses mature and accommodates more complex virtualization applications, it is necessary to allow VDI servers to provide GPU support to improve system performance.
GPU is often deployed as a separate device, but it can be integrated in a variety of different ways. The most common way is to install the GPU as an extended device such as a PCIE adapter card. Daily office computers usually use this method because there are many PCIE slots and easy to access, and the server can use powerful server -level products such as NVIDIA based on Kepler -based Grid K1 and K2 adapter. However, the server may not have enough PCIE slot to accommodate the GPU adapter. The GPU adapter is usually very large and is equipped with some cooling fans. The limited PCIE slot may be used for other extended devices such as multi -port network adapter or storage accelerator.
The other choice is to use external GPUs such as Cubix GPU-XPANDER to connect an independent GPU system that connects external and power supply by a simple, low-profile PCIE adapter. This method avoids the limited power supply of over occupation of limited servers and limits of PCIE slot space.
The third method is to directly integrate the GPU into the processor so that each CPU slot can access its own GPU. For example, Intel adds GPUs in the Xeon E3 processor and proposes a solution that improves transcoding performance to improve graphics performance. The RISC processor based on ARM architecture also adds GPUs to handle graphic tasks. Integrated GPUs may be the most effective way, because the power supply of the server does not need to be used to use the PCIe slot, but IT planners may need to wait for the technical upgrade to get the server that integrates the CPU/GPU. Some commercial systems are used to meet the needs of VDI hardware, but these systems are more pre -packaged instead of specially designed systems. An example is DELL's DVS simple device. The desktop virtualization solution (DVS) package is based on the Dell standard PowerEdge R720 or T620 server, and is bundled with the Citrix Xenserver or Microsoft Hyper-V and VDI management tools. According to reports, the device can support up to 129 users, and it is easy to deploy more devices to support more users.
It VDI devices include VMware's fast desktop device based on VMware Horizon View, Tangent's Vertex VDI device, and Pivot3's VSTAC VDI device.
The because DVS depends on the standard server, it does not make customized or specially designed to distinguish the traditional server. Such as N 1 redundant, automatic fault switching, load balancing, desktop configuration, and desktop mirror management can be implemented through software tools.
VDI instances support directly related to the computing resources, but VDI hardware requirements depend on the complexity of desktop mirrors and layered characteristics such as personalization and application virtualization. The above factors make the accurate amount of resources required for each desktop instance and the number of instances that a given server can support. This emphasizes the need for enterprises to test after a large -scale deployment of VDI, and in a well -planned POC project and a limited deployment environment (such as choosing a working group or department) for testing.