Often we are faced with situations where a server is required for a critical application, but during the course of procurement, as the proposal moves through the approval matrix, queries are raised as to why a desktop PC can’t be used as a server for a very obvious reason – Cost. It is important to consider cost effectiveness but at the same time it is pertinent to understand the very nature of an equipment and deliberate carefully on its pros and cons to make informed decisions.
It can be pretty difficult to explain the difference between the two to a non IT person, if we go in depth into its technicalities, so let’s use an analogy with SUV and sedan which could be of some help. Though a SUV and a sedan may have similar technical specifications like a engine capacity, power steering, ABS, EBD, climate control, seating capacity of 5 etc. still there is a vast difference in the manner a SUV is built and in the manner it handles on and off the road in comparison to a sedan. Similarly, though a desktop PC may have similar technical specifications and can run critical applications, there is a vast difference in the manner it would handle large amounts of data for extended periods of time.
1. Power supply
Unlike a regular desktop PC, a server is designed to operate 24x7. In order to achieve this purpose a server needs continuous supply of power. To ensure this, a server is equipped with a redundant set of power-supply (SMPS). This redundant set is designed to kick in and take over the duties without disruption, in the event of failure of the main power-supply. The power-supply of a server is also “hot swappable” meaning– a faulty power-supply can be replaced on the fly without shutting down the server. Power-supply of a server is also designed to handle power fluctuation/surges and spikes more efficiently than a desktop PC. It is not uncommon to come across servers which have never been shut down for several months.
 |
| Desktop PC |
2. Storage
Most desktop PCs come with a single hard disk. Though they can support at the most four hard disk drives, utility of such support is highly impractical as read write operation in such cases is not seamless, very slow and requires human intervention. Eg. Lets assume you are copying a large amount of data on a desktop PC which has four hard disks installed. Now if disk#1 reaches its capacity, the user will be notified that the disk is full, and the user will have no other option but abort the copy process and restart the same on disk#2. A server on the other hand supports several hard drives and can be configured to seamlessly perform read write operation over several hard disks using the on-board RAID controller. If you were to copy the data from the above example on to a server, the server would automatically copy the data to disk#2 as soon as disk#1 reaches its capacity. Not only does a server performs read write function seamlessly, its storage system is also fault tolerant -it can be configured to make copies of the data on one or multiple drives in such a way that data is not lost even if one or more drives fail. Like the power supply, hard disks of server are also hot swappable.
 |
| Front Panel of a server |
Processors, hard drives and other components generate large amounts of heat. Servers are equipped with several heat sinks and fans to keep these components from overheating. The cooling system is designed to minimize heat through efficient ventilation and maximize cooling by optimizing air circulation inside the cabinet. A desktop PC on the other hand has a very rudimentary cooling system.
 |
| Inside a server |
4. Cabinet
The cabinet holds the mother board, power supply, processor, disk drives, fans and other add-on cards. These components of a server are made of high quality material which make them relatively heavier than those used on a regular desktop PC, moreover, multiple quantities of these components are installed on a server. Hence, cabinet of server is built up of heavy gauge metal sheet to make it sturdy and strong to support the weight of all the components. A sturdy cabinet is also essential to protect its components from vibrations caused by multiple fans and disk drives installed on a server. To prevent deliberate switching off or resetting of server by unauthorized persons, the server cabinet is equipped with lockable switches.
 |
| Back Panel of a server |
5. Connectivity
Servers are equipped with at least two on-board Ethernet ports to ensure continuous and seamless network connectivity. Server can be provided connectivity through more than one network source/ISP, even if one of the sources fails the server can remain connected through the other source/s.
6. Scalability
Server unlike a desktop PC is highly scalable. A server is equipped with several additional bays, slots and sockets which has a lot of scope for upgradation or addition of components such as processors, RAM, hard disks, graphic cards, LAN cards etc. Desktop PCs have a very limited scope for upgradation.
7. Lifespan
A typical lifespan of a sever is 3 to 5 years, but timely upgrades and good maintenance practices can extend this range to more than 10 years. A well maintained desktop PC with average usage will perform at its optimal for about 2 to 3 years.
A desktop PC may be able to perform task of a server, but only for short periods of time, that too with below average performance. It can be used as a server during development, trial and testing phase of an application when there is negligible load. The performance of a desktop PC as a server during this phase should not be factored in while making decision on procurement of a server. As in the case of a SUV which is designed to take on off-road duties with ease, a server is specifically designed to host and run critical applications and process large amounts of data for extended periods of time, like a sedan which can never substitute a SUV, a desktop PC, can never substitute a server.
No comments:
Post a Comment