Ways in Which Fault Tolerance Can Be Added to a Network and Reasons for Use, and Disadvantages of Improvements - Assignment Example

Student’s Name Instructor’s Name Date Course Diagrams plus a maximum of 2500 words. Question 1: Redesigned network. T3 cable - Modem bank Key: CS- corporate staff workstation WS- Employees’ workstation. Question 2: Ways in which fault tolerance can be added to a network and reasons of use. A fault tolerant computer network ensures the system is still utilizable after faults because it has the means circumvent the manifestation and existence of faults. A computer network fault is created by a hardware failure in either one of the components or a long range of components that make up the network. A computer network is prone to a wide range of variables that can lead to its malfunctioning (Hanmer, 22-23). These variables include network adapters, power and power supply, network cable, hard disk, routers, hub, and switches. If a faulty power supply affects a computer system in a network, it shuts down, stopping all the executions. On the other hand, if the computer system’s adapter is faulty, it delinks the computer from the network. If the power supply to the network is faulty, the computer and devices making up a network are delinked from each other and from the network support devices like storage devices, hubs, and switches. All devices making up a computer network can fail at any point in the network’s life making it necessary to create a tolerant mechanism to counteract such failures (Sorin, 15-25). Failure of a network because of any of the reasons stated here will automatically have impacts on the business. If the failure includes the core areas of business or management, the effects can be very profound to the point of crippling the entire business. For the business to mitigate the risks of network faults, various fault tolerance methods are utilized. The basic methods of fault tolerance include the use of protected power supply and the disk arrays to create fault tolerance. Power failure encompasses the most significant point of failure of a network making it necessary to devise methods of proactively solving the problems of power supply and create power fault tolerance in the network. Disk array methods of fault tolerance seek to face the unreliable aspects of the hard disks. Currently, the quality of the hard disks produced by different manufacturers is high enough to guarantee freedom from intrinsic failures. However, they are still subject to degenerative failures. The degenerative failures of hard disks are primarily caused by the problems of power. However, these failures can be addressed via disk array methods as well as RAID systems. Fault tolerant RAID systems are supported by Windows 2000. The RAID system of fault tolerance is classified into several levels with the most popular levels being levels 1-5. RAID systems allow the computer users in the network to store information even after the failure of one of the hard disks (Bird and Mike, 177-83). Power supply protection. Fault tolerance can be added through power supply protection because of four types of power supply problems, namely voltage variations, long-term outages, short-term outages and failure of the local power supply. When adding an uninterrupted power supply mechanism on a network, it is advisable to examine the type of power supply failure as well as the network risks associated with the failure. The local power supply failure results from the servers’ moving parts and cannot be solved by uninterrupted power supply or backup generators but rather by a redundant power supply. Most of the current versions of servers contain a redundant power supply or an option for its addition to making the server tolerant to faults of power supply failure (Hanmer, 42-47). When planning to buy the servers, the organization should budget for servers with a redundant power supply or those with options that support its addition. If the plan is to use the existing servers that lack the option of a redundant power supply, it is advisable to include a spare power supply. Local power supply failure can also be circumvented through the replacement of power supplies of the critical hardware. Replacement of the power supply in short cycles increases the network’s tolerance to local power supply failures. Major voltage variations can easily cause computer/electronic failures on the network. Voltage variations in a network come in the form of brownouts, surges, spikes, and sags. Brownouts are reductions in voltage in order to protect the utility from overloading, most often by 5-20 percent. A network disruption because of brownouts can be avoided by the use of uninterrupted power supply (Hanmer, 42-47). However, if the brownouts are too frequent, the UPS will be damaged. It is recommendable to maintain high tolerance to brownouts through constant monitoring of the UPS’s health. However, the best protection from brownouts is achieved through a constant voltage transformer. On the other hand, power spikes are characterized by increased voltage because of such reasons as large motor starting as well as a lightning strike on a power line. If the cause of power surge is a lightning strike, the fault can be extensive. However, the fault tolerance in power spikes can be achieved through surge protectors. Surge protectors detect voltage excesses, protecting the computer network by rechanneling the excess power to a different electrical path (Bird and Mike, 17-24). Most UPS have power surge protectors, and it is advisable to locate their features to help create ultimate fault tolerance. Besides, tolerance to power spikes can be achieved via the use of the constant voltage transformer, suitably used in industrial setting to create fault tolerance. Protection from voltage sags addresses the problems of short-term voltage reductions. Mostly, power sags reduce the power line voltage to less than 100 volts causing the reboot of the servers. This problem can be addressed through constant voltage transformers as well as UPS. If a UPS is to be, it should be fitted with a battery power supply to address the problems of voltage sag. Power outages can be long-term or short-term and can cause a faulty network. Short-term power outages are limited to a few minutes and can result from either the internal or external events. Short-term power outages cause rebooting or catastrophic failure of the servers. Protection of the computers from the short-term outages is achievable through the use of UPS and a spike protection. Often, most short-term outages are followed by voltage spikes after their restoration. Besides, most power lines experience serial short-term power outages making it necessary to have a UPS and spike protection all the time. Long-term power outages last from few hours to several days, and results from such events as fires, storms, and earthquakes, amongst others (Hanmer, 49-517). Long-term power outages can be addressed via the use of an auxiliary generator and several UPSs depending on the extent of the power outage cause. The auxiliary generator should start automatically in the event of a long-term power outage to avoid any network fault. Disk arrays. The most likely failure of hardware in a computer network is a hard disk. Often, most hard disks experience failure in their first month of use. Hard disks are also prone to failure through the power-related degenerative and catastrophic changes (Rowe and Marsha, 33-39). However, these faults can be successfully addressed via RAID systems. RAID systems are available both for software and hardware. RAID implementation at the hardware level entails the use of RAID controllers while the software-based RAID implementation entails either the operating system or third-party add-ons. Comparatively, the hardware-based RID controllers are better than the software RAID because it supports on-the-fly arrays’ configuration, more RAID levels, and swap/spare drives as well as devoted reads’/writes’ caching. All RAID levels, except level 0, provide effective information storage in various hard disks to prevent the loss of information after disk failures (Bird and Mike, 17-24). Some RAID levels can provide protection from multiple hard disk failures. For the RAID system level to be effective on the systems, it is necessary such variables as the intended use, performance and fault tolerance. If the computer system is for the write-intensive application, fault tolerance cannot be achieved through software RAID or RAID-5. Such applications require the use of RAID-1, in duplex or mirror configuration to provide an adequate level of fault tolerance. If the application is read-intensive and uses sequential storing, it is advisable to use either RAID-3 or RAID-4 because they are effectiveness (Bird and Mike, 14). RAID-5 may be used in the read-intensive application. Before employing the RAID levels in a system, it is recommended to examine their levels of fault tolerance capabilities. For an RAID-1 duplex array, a complete failure leaves one copy of the data intact. However, after the failure of one disk the fault tolerance is lost. RAID-3, 4 and five arrays are associated with significant performance degradation after the destruction of one of the disks. However, if the failed disk is replaced fault tolerance is achieved through a complete rebuilding process. RID systems with multiple arrays also have multiple redundancy levels making them suitable for mission-critical applications. RAID levels’ performance is influenced by the intended use, but RAID-5 comprises the best compromise of any given situation. However, this compromise has the most widespread effects if the application is write-intensive (Arora, 53-57). A better option is to use RAID-1. Nonetheless, any level can provide substantial fault tolerance if there is the use of more small disks instead of a few large disks. This benefit comes from the parallel reading and writing because of the contributions of more drives to the array stripe. Other methods of ensuring fault tolerance in the computer network include clustering and distributed file system. In Windows 2000 servers, two different kinds of clustering can be used to improve the network’s fault tolerance. One of the clustering methods is the IP/TCP-based applications, and it provides a simple fault-tolerant application server through Network Load Balancing (Arora, 53-57). The Network Load Balancing allows IP/IP-based uses to disperse animatedly across a maximum of 32 servers. In this case, a failure of one of the servers results in the distribution of the connection and load to the remaining servers, creating a high-level fault-tolerant situation without the use of the shared or specialized device. In Network Load Balancing, every server can have specific capabilities and hardware and load balancing/failover occurs automatically without affecting the Network Load Balancing Service (Shooman, 62-69). The second method of adding fault tolerance through clustering is server clusters. Server clustering is dependent on the cluster nodes shared resources. The shared disk resource is a shared SCSI onto which every server in the cluster is linked (Arora, 53-57). Generally, the cluster nodes have identical capabilities and identical devices, but it is possible to create a clustered servers with dissimilar cluster nodes. Server clustering can provide a configurable and fault-tolerant environment to support the mission-critical applications and services. A Distributed file system simplifies the view of the network’s available storage. Besides, an appropriately configured distributed file system provides a good mechanism of fault tolerance. If the network systems distributed file system roots are configured on Windows 2000 domain controller, a flexible fault-tolerant distributed file system can be created. If the links of a distributed file system are replicated across a number of servers, a fault-tolerant file system with the ability to distribute the loads across the shares can be created. Question 3: Description of the design and advantages/disadvantages of the suggested improvements. The redesigned network has three line routers and an additional Wi-Fi router providing wireless access to the Internet through personal computers and mobile devices. The presence of a Wi-Fi router provides an opportunity for the organization to leverage the flexibility created by the wireless connection of the computers to the internet making it possible to handle some tasks seamlessly (Singh, 17-22). The company’s file server is redesigned to connect to the employees’ servers as well as the corporate servers to increase file sharing among the employees and the executive management. The employees’ and the corporate staff workstations are connected peer-to-peer through a common switch. The advantage of this peer to peer connection is its minimal resource requirements to connect the computers together. The workers can access all the files on the shared folders as long as the computer is connected to other workstations. This connection also means that work is not limited by the failure of one of the computers because each computer is independently connected to the switch. However, this connection has some security shortcomings because files within the network are only protected through passwords (Singh, 17-22). Besides the security concern, this form of connection does not allow an efficient running of the computers because it supports optimum functionality in the range of two to eight computers. Moreover, connection setbacks can cause problem accessing some files. On the other hand, the corporate staff workstations are connected to each in a client-server network. In this network, there is centralization of resources, access, and data security because computer interaction is achieved through a central server. Besides, any workstation can undergo upgrade without causing unnecessary network downtimes in the corporate offices. Moreover, this connection allows easy integration of new technology because of its high flexibility. Furthermore, the components of the client-server connection of the corporate staff’s network are interoperable. However, the connection is expensive to achieve because of the high start-up cost. Additionally, the problems of the server can render all the workstations unusable until the server is surfaced. The redesigned network has additional servers, switches and routers Each switch is linked to at least two routers making them fault tolerant in case of the failure of any of the routers. The switches are also interlinked to, and every workstation is independently connected to the switch reducing their interdependence. The network also has access to multiple UPSs to create tolerance to power-related network faults. Work cited. Arora, Aakash. Design and Performance Analysis of Fault Tolerant Ttcan Systems. , 2005. Computer file. Bird, Drew, and Mike Harwood. Network+. Indianapolis, Ind: Que, 2005. Print. Hanmer, Robert S. Patterns for Fault Tolerant Software. Chichester: John Wiley & Sons, 2007. Print. Rowe, Stanford H, and Marsha L. Schuh. Computer Networking. Upper Saddle River, NJ: Pearson/Prentice Hall, 2005. Print. Singh, Vishnu P. Computer Networking Course. New Delhi: Computech Publications, 2010. Internet resource. Shooman, Martin L. Reliability of Computer Systems and Networks: Fault Tolerance, Analysis and Design. New York: Wiley-Interscience, 2002. Internet resource. Sorin, Daniel J. Fault Tolerant Computer Architecture. San Rafael, Calif. (1537 Fourth Street, San Rafael, CA 94901 USA: Morgan & Claypool Publishers, 2009. Internet resource. Read More