Infrastructure

Single or multiple tower servers

Tower servers can seem dated and look more like desktop units than servers, but these servers can still serve a purpose. Generally speaking, if you need a lot of servers, you will probably not use tower servers. They can take up a lot of space and are tough to physically manage since you can't easily stack them on one another. In some cases, as businesses grow and move to rack servers, conversion kits can be purchased to turn a tower server into a rack-mount server. This depends upon the age of the original server and can sometimes be more cost-effective to replace the old servers when they reach the end of their life cycle.

As implied, tower servers are probably more likely to be used in smaller environments, although you might find them used in larger organisations. For example, they can be useful as a branch office server.

Tower servers are generally at the lower end of the price scale, although, unlike desktops, they can usually be expanded a large amount and become really expensive. Another advantage of tower servers is that they can fairly easily be moved from one location in an office to another.

Tower servers take up a lot of space and usually require individual monitors, keyboards, and mice or a KVM device (keyboard, video, mouse device that allows them to be managed with a single set of equipment).

Additionally, cabling can become really messy, especially if you have a lot of network adapters and other I/O needs.

Rack mounted servers

If you run a business of any reasonable size, you will probably want to use industry standard 19" wide rack servers. As indicated, rack servers are permanently mounted in a rack or cabinet, they can take up varying amounts of space in the rack from 1U to 8U (rack servers are sized in "U", which is a single 1.75" rack unit). Generally, the larger the server, the more expansion opportunities are available.

Rack servers are extremely common and make their home inside these racks along with other critical datacentre equipment such as switches, storage arrays, UPS (uninterruptable power supplies or backup battery units).

Rack servers make it easy to keep things neat and orderly since most racks include cable management of some kind. However, you still need a lot of cabling to make everything work.

Similar to tower servers, rack servers often need a KVM device in order to be managed.

Rack servers are very expandable; some include space for 12 or more disks right in the chassis and support for four or more processors, each with multiple cores. In addition, many rack servers support huge amounts of RAM, so these devices can be computing powerhouses.

Blade servers

A blade server is a stripped-down server in a modular design optimised to minimize the use of physical space and energy.

A standard rack-mount server can function with (minimally) a power cord and network cable. Blade servers have many components removed to save space, minimise power consumption and other considerations, while still having the main functional components to be considered a computer.

Unlike a rack-mount server, a blade server needs an enclosure. A blade enclosure, which can hold multiple blade servers, provides services such as power, cooling, networking, various interconnects and management. Together, blades and the blade enclosure form a blade system.

Originally, buying individual blade servers could mean trading expansion possibilities for compactness. Although this is still true to some extent, today's blade servers can be quite powerful.

There is still some truth to the fact that blade servers have expansion challenges when compared to the tower or rack-based options. For example, most tower servers have pretty significant expansion options (expansion slots, disk drives etc.). Many blade servers are limited to two to four internal hard drives, although organisations that use blade servers are likely to have shared storage of some kind backing the blade system.

Furthermore, when it comes to I/O expansion options, blade servers are limited by their lack of expansion slots. Some blade servers have PCI or PCI-E expansion slots, but for most blade servers, expansion is achieved through the use of specially designed expansion cards such as multiple Gigabit Ethernet adapters, Fibre Channel adapters etc.

When raw computing power and server density is the key driver, blade servers meet the need. For example, you can easily fit a 10U blade chassis into a rack that can support up to 16 servers. So, each server uses the equivalent of 0.625U of rack space. The blade chassis might have inbuilt gigabit Ethernet switches, Fibre Channel switches, KVM device etc. so there is additional rack space savings since there is no need to rack mount these devices to support different connectivity and management options.

A blade environment generally has much less cabling than tower or rack environments since a lot of the connectivity is handled internally and you will end up with a neater server room as a result.

Adding a new server consists of simply sliding it into an available slot in the chassis. There is no need to rack a new server and deal with mountings, new cabling etc.

The combination of this small size and high density makes heat dissipation a challenge. Blade chassis can put out a lot of heat so ideally, the server room should be air-conditioned.

From a cost perspective, blade servers require some initial infrastructure, such as the chassis, so the upfront cost is often higher than for servers of other types.

Virtual servers

As the name suggests, a virtual server is an operating system environment that is installed as software on one of the hardware servers mentioned above. It doesn't actually exist as a piece of hardware in its own right. In order to implement virtual servers, you need at least one medium to high end hardware server which will be capable of running multiple instances of virtual servers. Frequently, the hardware server that is running the virtual server(s) is known as the host and the virtual server is known as the guest. The virtual server imitates dedicated hardware and the end user or client software has the same experience accessing a virtual server as they would if accessing dedicated hardware.

Using virtual technology allows an organisation to implement multiple servers to suit their needs without the physical space penalty often associated with that. A well-planned environment using virtual servers is usually more cost-efficient than using dedicated hardware.

Cloud-based servers

Cloud based computing is usually based on the virtualisation technology mentioned above and is experiencing exponential growth, with companies of all sizes adapting to this new technology. Industry experts believe that this trend will only continue to grow and develop even further in the coming years. While cloud computing is undoubtedly beneficial for mid-size to large companies, it is not without its downsides, especially for smaller businesses.

If used properly and to the extent necessary, working with servers and data in the cloud can vastly benefit all types of businesses.

Storing information in the cloud gives you almost unlimited storage and computing capacity. Cloud based servers are highly scalable so you can start off with one or more small severs, add more processors, disk space or additional servers as you need, you can even have some servers off in out of hours periods to reduce costs.

Cloud based servers can be sited in data warehouses around the globe offering the ultimate in fault tolerance. Since your data is stored in the cloud, backing it up and restoring data is relatively much easier than storing the same on a physical device. Backups are easily stored in a different physical location giving ultimate protection and ease of disaster recovery.

Cloud-based servers can be set up so you can access data from anywhere you have an Internet connection. This convenient feature lets you move beyond time zone and geographic location issues.

Deploying cloud-based servers can be relatively quick compared to deploying physical servers. Of course, the amount of time taken here will depend on the exact kind of technology that you need for your business.

In spite of its many benefits mentioned above, deploying cloud-based servers may not necessarily be the solution for everyone as there can also be some disadvantages. For example, if you opt for a cloud only solution and for any reason you have no internet connection (office internet connection goes down, mobile based connection and there is no signal), you will have no access to the servers or the data contained upon them. In any event, you will need a very good Internet connection to be access the server at all times.

Raid arrays

RAID is an acronym for Redundant Array of Independent Disks. This is a data storage technology that combines multiple disk drives into one or more logical units for the purposes of data redundancy or performance improvement. Data is distributed across the drives in one of several ways depending upon the design goal. RAID is a cost-effective solution often used for providing fault tolerance on servers to negate the effect of a single disk failure. RAID arrays can be mounted physically in a server and are part of the computer subsystem or they can be housed in a separate unit, either standalone or rack-mounted.

NAS

NAS is an acronym for Network Attached Storage. This is a file-storage device that is accessed through a network allowing multiple systems to access the same data. Investment is protected with NAS solutions as more NAS appliances can be added to the network as required. A NAS device can be either standalone or rack-mounted.

These devices are in fact usually just another computer on your network with their own processor(s), memory and a purpose-built operating system that has the dedicated function of storing and supplying vast amounts of data. They frequently have multiple network ports (Gigabit or 10 Gigabit), SSD cache acceleration and almost limitless expansion capacity.