Six major dangers of blade server

Wikipedia defines http://www.aliyun.com/zixun/aggregation/13737.html "> Blade server as: Streamlined, modular design optimized to minimize physical space. Compared to standard rack servers, which have to work through power lines and network cables, blades have many components that can be removed, resulting in space savings, minimizing energy consumption, and other advantages, while retaining all the necessary functional components, so they can still be identified as computers.

It is important to define blades as it is becoming more common, even in the second-hand market, because dealers may deliberately confuse the concept of blades with standard 1U and 2U rack-type servers to fool novice customers. The enterprise is to be reminded that the blade is a special hardware category, it needs to use a chassis, rather than a simple "small" server, the blade chassis with a variety of shared components, such as power supply and remote management console, thereby reducing the necessary components of each individual blade server. Therefore, the risk of blade server, can not be belittled.

The first danger is the renewal of blade technology.

Unlike traditional racks, which have barely changed in decades, Blade chassis technology is up to date and often varies. Since the advent of the 2001, blade servers have been in a hurry for several generations, each generation will require replacement of the chassis to support the new blade server. If you buy a few servers, this is a big risk, your return on investment or even not the cost of replacing the chassis. The rate of change will gradually decrease with the maturity of technology, but the risk still exists.

The second danger is the cost of blades, especially the blade chassis

Blade enclosures are generally very expensive, even if the blade itself is cheaper than a rack-less server. Quick glance at the major blade suppliers of the product quotes, the chassis is generally about 5000 U.S. dollars, generally accommodating up to 8 blade servers. Each blade is about 500 dollars cheaper than a rack-like server with the same specifications. This means that even if a blade chassis is fully configured with 8 blades, it will cost more than 1000 dollars in price, and each additional blade slot will be thanks to an extra 500 dollars. The trouble is that the blade chassis usually holds 8 to 16 blade servers and needs to be purchased in advance. If you need enough servers to work for you, then there is no problem buying a matching chassis, but if you need only one server now, the investment will be a bit too big.

The third danger is shared backplane and other important components

Although the blade chassis has generally built up a lot of redundancy through amazing design, it still has to consider the problem of a single point of failure. For a rack-like server you can add redundancy by adding additional servers, and you can configure it to your requirements. But for blade servers, you have to add more chassis to achieve the same level of reliability.

Repeat it again! For large enterprises This may not be enough, but for small businesses, it can be a pain in the neck if you suddenly need to buy another chassis to complete redundancy.

The fourth danger is flexibility

Small IT vendors generally do not easily handle servers. If a small business with three servers buys a new server, it usually arranges the old server into another role-perhaps in another office. But the old blade server can only be placed in the slot where it can match the blade chassis. This usually results in the blade losing the opportunity cost at the end of the lifecycle, and this part is often overlooked in the cost analysis of the blades. Without a suitable blade slot, the old blades are more likely to be forgotten or discarded, rather than redeployed.

The fifth danger is that the manufacturer locks

Traditional rack technology has nothing to do with manufacturers. Most service providers mix and match a variety of servers, power supplies, routing, switch monitoring devices and other devices in the rack of the computer room. The blade server must use a specific vendor. This may not be a problem for large enterprises, but for small servers with limited server size, there is no way to abandon different vendors and technologies.

Suppose a small service provider wants to apply the Intel Itanium processor's HP Integrity Blade server, they will buy the blade chassis and start using it. In three years, they will also buy software that runs on Sun UltraSPARC and IBM power processors, each of which requires its own branded blade chassis. This will greatly increase the risk of a small service provider if the chassis slot is not full.

The sixth danger is the high cost of storage

The storage options available today include San, Nas, Das, and so on. Service providers of all sizes are rapidly moving closer to SAN and NAS because they provide enough networked storage to mitigate the risk of storage for many blade servers. However, many service providers do not know who is the San and blade who helped. So a comprehensive look at servers and storage is critical.

A typical blade server can only hold one or two 2.5-inch SAS or SATA drives. This is far below the potential storage space provided by a typical rack server. In a popular 2U rack, 8 to 16 drives are common, and some use 3.5-inch drives instead of 2.5 inches. There is a popular very cost-effective 2U server that can provide 28TB of low cost storage capabilities. And you can't use this type of storage for the blade chassis, but the only reason is that the blade chassis does not provide the space for the local drive. The owner of the blade can only be forced to use the smallest direct additional storage-even if DAS can provide better performance and cost.

To make up for this, blade vendors provide storage blades-small, low capacity SAN devices-directly into the blade chassis. But these devices are usually low performance, often only six drives, and are too expensive compared to other storage methods.

Obviously, buying a blade chassis doesn't mean you've given up on using a rack-style server and you can continue to mix and match. But for small businesses, the use of blade architecture often leads to a non-stop purchase of blades to fill the chassis vacancy path.

Is the blade losing its chance? Most of the risk of blade servers means potential losses. Compared to large enterprises, the advantage of SMEs lies in flexibility and agility. And the blade stands on the opposite side of agility. They require large upfront deployments and the planning of the underlying architecture, with a large demand for technology, physics, and space. Even if you don't think deploying a blade will have any impact on your future business, that doesn't mean you won't miss the opportunity. Because the blade lacks flexibility, it is difficult to adapt to changing conditions.

But this is not to say that the blade server is meaningless. Because of shared system components, blades are much lower in power consumption than traditional servers. The blades are easy to transport, so as long as the chassis is empty, it is easy to move. Almost anyone can easily move a blade server.

If combined with the San underlying architecture, blade servers are very useful in virtual environments. This combination maximizes the cost and flexibility advantages-and, of course, for large businesses. For small and medium-sized Enterprises, the application of virtualization technology means to reduce the required server, so the blade architecture is difficult to take advantage of the previous need for more than 10 or more servers to complete the work, today, only three or four can do better.