A network is defined by Noergaard (2005) as a collection of two or more connected devices that can send or receive data. He further adds that the relationship between these connected devices in a network determines the total architecture. The purpose of this report is to understand the networked applications architectures. Two networking architectures are covered here; the client/server architecture and peer-to-peer architecture. The report first gives an outline of these two architectures and their advantages and disadvantages.
It is seen that the industry divides its products on the concept of client-server model or peer-to-peer model. Some applications perform better with the former model while others perform better with the later model (Klemt 2005), 1992). Based on the understanding of this report, two separate scenarios for a basic networking application will be analysed. The idea will be to judge the type of networking architecture to be used for each scenario. Finally based on the entire analysis, a conclusion will be drawn as to which resource has to be used in what situation.
Part 1: Review of Client/Server and P2P application architecture Client/Server architecture Client-network computing is one of the most basic paradigms of distributed computing, and this architecture is based on the paradigm as described by Jalote (2005). This architecture has a minimum of three components – a client, a server and a network that connects these two systems. This is a two level architecture where a client requests a service and the server provides it (Singh 2005). The server is usually accessed through a remote procedure called RPC, request of which is defined by their signature.
The processing of the request is internal to the server and is transparent to the client. The client and server may be on different or same systems Advantages of client/server architecture Client server architecture using RPCs has been used for over a decade for system wide exchange computing (Singh 2005). The main reason for this is the high level of reliability this architecture provides. It is also found that the cost of the entire system over a period of time is less than that used by other architectures.
In addition, the architecture also enhances safety since any of its components can be removed without a major impact on the other system. This architecture leads to a better distribution of resources through the system. This architecture also leads to increased productivity because the front and back end distribution of processes is optimised (Jalote, 2005). Disadvantages of client/server architecture The main disadvantage of the client/server architecture is that there are multiple databases in the client/server environment, making the management and synchronizing of the data shared between them difficult.
This also leads to a difficulty in software distribution and configuration management program, as the system is distributed. The cost of maintenance may also increase because of the different networks being inter-networked. P2P architecture Peer-to-peer or P2P architecture is another way of managing distributed applications. In this method, the individual nodes have similar and comparable roles, unlike the client/server systems. In this architecture, the nodes are given roles such that both can act as client and server. This type of connection requires at least two computers connected to each other (Reid, 2004).
In order to function correctly each node of the system should know the address of at least another node. Each node can directly contact the other node of the system for using its resources, without going trough the server (Heathcote & Langfield, 2004) Advantages of P2P Architecture This type of communication exists since the time of old
Also the combination of processing power of multiple nodes in a network greatly increases the efficiency of the system in terms of combined processing power and storage when compared with a central server. P2P systems are hence extremely autonomous and can secure anonymity. They can be generally deployed in an ad hoc system without requiring central management or control. Finally, unlike the client/server systems there is no need for a separate network operating system (Reid, 2004). Disadvantages of P2P Architecture The manageability of peer-to-peer systems is much harder, due to the extremely distributed nature of the system.
Security is also not centrally controlled and the users need to remember different passwords for accessing the resources located on different nodes on the network (Porter, 2006). Facilities such as backup are the responsibilities of the individual user unlike the client/server systems where the server can schedule them. This system works best for lesser number of systems in a network, usually under 10 (Reid, 2004). Part 2: The architecture of Mr. Robertson’s home network Network Architecture The architecture which I would suggest for Mr. Robertson’s home would be a peer—to-peer network.
There is no necessity for a client/server model in this case. Also no type of hub system is required, since the application is very simple Justification As is already explained in the section above, a peer-to peer network is used for ease of applications without extra overload. The system works best for systems which having less than 10 systems. As Mr. Robertson’s system uses only two computers, the system is ideal for this purpose. In addition to this, the home network does not require the additional overhead of a server so a pure peer-to-peer system, one without any server (Subramanian & Goodman, 2005), can be used.
The application mainly used is the file sharing which can be easily done. Also for using the internet, there is no necessity for the system to have simultaneous connections as is already mentioned in the example. While, to most intents and purposes, a printer application requires the use of a server so as to manage the print request being sent from multiple systems, this is clearly not the case here. The network should merely have the capability to receive print jobs from both the systems a simultaneous operational capability is not really required, as the application would not be too critical.
In the worst case, the print job could be cancelled and given again. Finally, the use of a client server system would in addition to setting a server, also need additional components and software like an operating system, which is not required in this case, as there is no need for tasks like allocation resources, backups etc. Also, Mr Robertson gets the advantage of removing any of the two systems from the network at anytime, and the second component would work just as well, as neither of the system is a server. Components for the home network In addition to the 2 computers and printers purchased by Mr.
Robertson, he needs very few components to get his system up an running. He would need the cable modem, a twisted cable pair, media converter for running the internet through cable, a splitter and a print sharer. There is no necessity for a hub for such a system, which is necessary only if there are more than two computers on a network. The splitter would divide the internet connection to the two systems which will the function of the print sharer also. A wireless NIC or a wired one can be used to share the network. Only single numbers of components, except the connecting cables, are needed because there are only two computers.
Topology of the home network The figure above shows the basic topology of the network. As explained above, there is no need for a hub for just 2 systems. The printer can be connected either via a print sharer or the software can be used to do that by accessing the network. Here, each computer can access the network printer. In addition, the cable connection that comes can be split into two using a splitter. This type of networking requires basic minimum of components and can be used for connecting systems on a home network. This system is extendable to more than two systems.
The only difference would be that in that case, a hub would be used between the two systems instead of just a cable. Also the cable type, which is twisted in this case, might be changed for more than two computer systems, because of load factors. In this case also, there will not be any necessity for a server, as the printer tasks will be shared, but not scheduled or managed by an external system. Part 3: The architecture of the LAN of BC Inc. Network Architecture For the company’s LAN, a client server system is a must. There are only 16 systems, on the LAN would more than fulfil the application necessity.
There are multiple applications needed for this system. There is the printer service which is multiple in numbers and locations. In addition, there is the obvious necessity for an email and web server for the business applications. All this necessitates are need for the use of a client/server model. Justification In many of the situations given above the peer-to-peer network fails to have any use. First and foremost is the number of systems. The P2P system is only suitable if the number of systems is less than 10, as mentioned earlier. Also it is not just the number of system that is a problem, since it is not a very huge network.
Office applications as a rule need more features and functionality than a home based server, and so the load on the network contributed by each system is also a lot. In addition to this there is he additional necessity for security and backups as compared to the home networks. People usually prefer manual backups on their home networks, but almost all the companies insist on having scheduled backups for the protection of data. In addition, the systems also need to have security updates with a controlled installation, which is not necessary for a home network, whose security necessities extend maximum to having a basic firewall.
As in earlier case, due to data sensitivity, companies prefer to have all of these controlled rather than leaving it to the discretion of the office employees. There is an obvious burden due to the higher cost of installation and maintenance of the system. But this is more than compensated by the quality of security and safety levels given by such a system. Also since, there are only 16 employees in the system, it is logical to have the email server as the same as the network server. This is also true for the web server applications. However, here there is an assumption being taken.
It is assumed that the company business is a small size, based on the number of people working there. This may very well not be the case. In addition to this, the web sever installation might be mandatory on an external system, because of security reasons. For instance, the company might not want its intranet lined with its webserver to provide more security. This case might necessitate the need for a separate web server, though for this report, a single sever is being considered for now. Physical Components In addition to the computer systems and printers, the company needs many additional systems to make a network.
First and the most important of this is the server system, which needs be high ended, with high processing power and speed. This is that the system can easily act as a intranet server, a mail server, and also the web server. In addition to this a router is essential for connecting to the internet via the DSL modem. Also there are multiple numbers of cables, based on the systems connected to the network. Since the system is connected via the Ethernet here, the internet being a DSL based one; the corresponding cables are also required.
The printer is connected to the server, which takes care of the scheduling the print jobs. The other servers, as mentioned earlier are situated on the main server itself. The figure above shows the scheme of the connection diagram. The above diagram is a very simplistic way of representation of the office network. As mentioned above, the server would be common from email and web applications, and also take care of the intra networking scheme. In addition, the various printers and DSL modem would be added to this server only.
The server would need a network operating system and the main security firewalls and the protection schemes like the antivirus would be installed here. Ethernet would be the mode of interconnection in the network. The system can be extended when a new web server is required, which is separate from the present server. There can be another server attached to the present one which would take care of the business related websites etc. The connection would be restricted so that only the purchase related or feedback related information is passed to the main server, which would then direct this information to the necessary client system.
Conclusion In the report mentioned above, there were two types of networking architecture which were analysed. One was the client server architecture, which is one of the oldest networking schemes available. The system finds its use in complex networks having many systems, and necessity for multi-file sharing applications, common or shared database and also security. All these are provided by this system, which the case seen by the company example is being taken here.
The operations are very complex with multi-printers and their multiple locations, the necessary for common access to the internet, and also the need for servers for websites and emails. Also the network needed security and will be needed managed access permission for files stored on the network. The restricted type of entry for a network is possible only for a client and server type of architecture, which is what was proposed for this case. The company being an accounting firm would also need internal data restrictions and security issues, which can only be tacked by this architecture.
In contrast, the home networking system is a very simple operation, which needs to have a mere connection to the internet and the use of a printer. There are no constraints of security or access restrictions on this systems and the main application need is a basic file share operation. For such a system a basic, peer-to-peer architecture is sufficient. There is no need for hubs also as there are only two systems. The network does not need any server based system which is too much of overkill here. Hence, using these two examples, the two types of networking architectures are explained here.