Choosing a right server is not an easy task especially when there are thousands of web hosting providers all offering almost identical web hosting packages. It is therefore very important to consider a number of factors before choosing on the type of server to be used. Many companies will of several megabytes but one question to asked is whether the big space offered is really needed.
Depending on the capacity of an organization, various server hardware systems can be used. This depends on the choice of variety, the availability of relative resources to be used in each case, related costs, volume of transactions, size and accuracy of out put and the structure of information flow system within such an organization. Therefore, for different server systems, different information flow system can be achieved with each having various disadvantages and advantages.
Before I decide for Howard Rubber and plastic company on the type of server to be picked it is important to look for various advantages and disadvantages each of the servers have in comparison to the other. First of all it is a medium sized company comprising of 350 employees so it means that the server chosen should not be very expensive. The many departments the company has such as sales, administration, quality control, manufacturing and warehouse and shopping will force the company to make an appropriate choice of the server in terms of hardware’s cost, software cost, benefits and drawbacks. Let us now look at each of these three servers the company want to choose in order to make a good judgment on which one to pick.
Blade servers are sometimes known as a poor man’s supercomputer. They are just very simple devices involving the use of self-contained system of computer servers that are used in ensuring high density. Each blade in a blade server is an independent server. A blade normally has its own processor, hard drive and operating system. The moment the blade is plugged into the chassis, it gets to share power, optical drives, parts and switches with all other blades. The most important thing about blade server is that it shares resources. The cost of blade server is not all that expansive. It is estimated that the cost of a 64-bit blade server can be as little as $1,500. It therefore means that the company manager can save some money when using the system- because of the cheap initial cost. Cabling is simpler and this also cuts down the labor costs. There is also reduced cables clutter, which also saves space to be used1.
With the department the following advantages will be realized. One is that most blades are hot swappable meaning that if the manager is experiencing a traffic surge, he may move some blades from one department over his web server, temporarily, with no downtime. A gain there is no downtime should the blades break. What can be done when a blade breaks is that it can just be pulled out and a new one is plugged in during replacement. Blade servers are therefore economical and useful in a limited space.
Grid server on the other hand comprise of a collection of computing resources connected through a network along with grid middleware that aggregates these server. This server is not cheap in terms of power and also expensive in terms of cost but it has some benefits over other servers. It enables organizations to harness their computing resources more effectively by applying a more usable computational power to their most important tasks. Organizations also access the computer grid as a single computational resource thereby freeing them from the need to seek out appropriate system to run their jobs and applications. Grid server system also helps to reduce time taken to move goods to the market. By using a grid server the company always make it easier to improve the quality of its products in the market place. It encourages innovation therefore moving the company to a better position in a global market. Grid server also leads to better utilization of resources. It helps to protect IT investments by ensuring that they are optimally utilized. It also ensures increased productivity. It enables work to be computed more quickly since workers can focus on design and development rather than wasting time hunting for resources and waiting for their computer jobs to provide results2.
The grid servers also allow for scalability and flexibility in the organization. Different components can be up dated independently and additional resources can be added as needed, reducing one-time expenses. Grid computing also provides computing power where it is needed.
Before looking cluster servers in terms of costs, benefits and draw backs, let us look at the meaning. From the name cluster it means the aggregation of multiple stand-alone computers linked together by software and networking technologies, to create a unified system. There are basically two types of cluster serves such as high performance computing (HPC). The HPC is made up of markets traditionally serviced by super computers for application requiring greater computational power than a single computer can provide. The other category is the enterprise or High Availability (HA) with automatic failover, load balancy, redundancy and other features that provide high reliability for the data center. However many HPC clusters also incorporate some of the features of HA clusters. Mode of Application-varies between and within each of these systems.For example HPC clusters are normally used by engineers, scientists and researchers whose work generally demands the highest levels of computational analysis, modeling and simulations. Coming up with a highly efficient and usable cluster is just too expensive. It basically demands more than simply linking servers together with high-speed networks. The whole process of clustering is very complex and complicated. This implies that the company intending to use the cluster servers has to part with some money before the whole system is installed. Both software and hard ware costs are just too high for small and medium sized companies. Despite the costs, a number of benefits are brought/ realized when using server clusters because of the popularity it is gaining for both high performance and high availability applications. For instance, Linux provides costs advantages, greater flexibility for programming and fewer licensing requirements. Security is also guaranteed3.
Cluster server is a kind of software designed to allow servers to work together as one machine to provide failover and increased availability of applications. High performance computing cluster technology is used in supercomputers. It does not require shared disk as it only does calculations. Large companies need plenty of server horsepower to support thousands of internal employees and millions of daily visitors over the Internet. Clustering provide for higher business availability of search and indexing services. For instance a situation of malfunction of a server in the cluster, the system as a whole will continue to operate. Search results are being received and processed from the rest of working services. However some documents will be missing in the search results from the malfunctioning server. This can bring a big problem to the company. The only advantage is that the server will continue to operate even before the problem is fixed. This ensures symmetric data distribution among all cluster nodes.
Asymmetrical indexing methods of servers also bring some benefits. The method can be used when large portion of data is to be updated relatively less frequently than other small portion of data. Largest data set are indexed on one cluster and small data set with frequent updates indexed on the other cluster server. This method is good in economizing resources because all frequent updates are being done for the database of rather small size that provides higher total performance of the system and less frequent disk equipment. It also leads to higher reliability in the environment. This is because there are fewer chances to break something in small data set that is being always changed frequently. The server clusters are not cheap to install but they reduce the power usage hence low spending by the company since high power means high complexity. Therefore hardware and software bundling in server appliances help several companies to save a lot of cost. Server appliances consume less power than traditional servers. The only difference the server cluster has is that they can only handle a small range of jobs. It therefore means that a company, which is having hundreds of employees, like Howard Rubber and Plastic Company with 350 employees cannot effectively use it.
Coming back to Howard Rubber and plastic company, I would suggest the blade servers to be used. The first reason why I would prefer the blade server is that it is less expensive as compared to other servers. This is why most people name it a poor man’s supercomputer. Both the hardware and software costs are not high. If take for example cluster servers, several computers has to be linked together.4 This means that the cost of installation is very high for the company. 350 employees are not many so there is no need for it to use a grid server or cluster servers.
Comparing the costs of the various servers discussed, I believe the cost of a blade server is the cheapest or lowest. Cabling of blade servers the cheapest or simpler thereby cutting down the labor costs. Lastly blades server does not consume a lot of power therefore the company will have to save a lot when it uses a blade server.
Work cited
- Jim Hoskins. Exploring IBM server & Storage Technology. A laymen’s Guide is the IBA. London: Maximum Press, 2005, p. 86-88.
- Pawel Plaszczak and Richard Wellener. Grid Computing: The Savvy Manager’s Guide. London: Elseviers, 2006.
- William Schimed and Robert Shimonski. MCSA/MCSE Managing and Maintaining a Windows Sener, 2003 Environment for in – London: Elsevier, 2003, pp 75-77.
- Nathan, J. Muller. Networking A to Z, New York: McGraw-Hill professional, 2000, pp 289-291.