|← WAN Design||Relationship to Technology →|
Buy custom The Network Upgrade essay
This paper concerns a number of network issues experienced by a law firm. The company occupies four floors in the business center, operating a network with four respective segments. The general problem is an increased network latency, accompanied by the need to provide a sufficient security to the Accounting department. In addition, a new videoconferencing system is about to be implemented, which requires a proper design and testing.
The main reason for the increased latency is a network congestion. As all nodes on each floor have to share the resources of a single segment, the network throughput has reached its limit with the increased performance of PCs and other network devices. According to Karris (2009), “a network segment is a part of an Ethernet or other network, on which all message traffic is common to all nodes, i.e. it is broadcast from one node on the segment and received by all others” (p. 5-13). The majority of all network nodes in the company are connected to hubs, the network devices that emulate the 10Base2 Ethernet bus. Regardless of the physical star topology, only one network node can transmit at a time. If two or more hosts would start a simultaneous transmission, a network collision will occur, forcing all nodes to stop the communication and wait for some time before the next transmission attempt. These network features, combined with the relatively low data transmission speed (10 Mbps), result in a significant network latency. Therefore, the task is to provide the network nodes with a dedicated media instead of a shared one, increasing the network speed along the way.
There is an assumption that the cabling is done using UTP Cat 5, which is sufficient for the 100 Mbps network speed (White, 2012, p. 403). In this case, the hubs could be just replaced with the switches (Diagram 2). Instead of forwarding the traffic to all ports, the switch distributes the network packets based on the destination MAC (Media Access Control) address. It would be a good idea to use 100 Mbps switches with 1Gbps optical uplinks to the datacenter. Once such reconfiguration is performed, there will be no issues with the network latency. Moreover, every node will be provided with the dedicated 100Mbps access to the network resources.
The network security will benefit from a new switched environment too. In a shared segment, any node could potentially intercept the network traffic using the “packet sniffer” applications (Dean, 2009, p. 673). It would be no longer possible in a switched environment. In addition, the network should be split into several subnets. A router should be installed between the subnets in order to forward the traffic in accordance with the routing policies. Such policies could include the packet filtering, a security feature that would protect the network resources from the unauthorized access. As an extra measure, firewalls can be installed between the subnets operating the most sensitive data and the rest of the network. Most commonly, the firewalls are used only to guard the outer company’s perimeter.
In addition to the network upgrade, the company wants to implement a video conferencing system, perceiving it as an essential component of its business. The system will be used as a mean of communication with two new satellite offices and must be operational in six months. Apart from the architecture and design of the videoconferencing solution, it will require a thorough testing prior to its operational use. There are two options available for the testing program. The Prototype testing implies building an isolated environment that would resemble the conditions of the further system’s use. The alternative is a Pilot testing mode, when the system is tested in a real environment with certain restrictions.
The recommendations for this scenario will be in favor of the Pilot testing. First, the Prototype environment could not model all conditions of the system’s operational use. The communication links to the satellite offices will be slower compared to the LAN connections. It would be possible to emulate the slow connection using a traffic shaper, but in this case the QoS (Quality of Service) mechanism will not work as required in the real-life scenario. Second, it would be difficult to simulate a real network delay, which surely will differ in the WAN environment. Therefore, all the adjustments performed in the Prototype mode would be useless in the real system operation. In case with the Pilot testing, the only downside is a restricted testing time. However, as there is a six-month project timeframe, the WAN environment could be built a little earlier to allow at least a few weeks of the videoconference solution testing.
Finally, all servers, workstations and other network nodes should use a unified naming convention within a company’s namespace. According to Dean (2009), “the term namespace refers to the database of Internet IP addresses and their associated names” (p. 164). Taking into account the potential necessity to access these resources from outside the company network, the naming convention should be in agreeance with the DNS (Domain name system) structure. This structure is based on the several top-level domain (TLD) names, such as .org, .gov, or .com. Assuming that the company’s domain name is legalgeniuses.com, all network nodes should use it as a suffix. There might be a number of sub-domains too, such as hr.legalgeniuses.com or accounting.legalgeniuses.com. Subsequently, a printer in the CIOs office could be named bossprinter.it.legalgeniuses.com.