LO3

A server is normally considered to be a powerful computer or device which serves a particular need for a network. It can also be a piece of software that operates on a network device. The most common servers are the following:

File- A file server is useful for businesses or individuals who need to hold and manage large numbers of files. It stores files, indexes them, remembers where they are to retrieve them and takes responsibility for security of the files, ensuring those who are entitled to read them can do so and only those who have the rights can edit or delete them.

Print- Print servers manage the printing of documents. They control printers on the network, allowing a group or groups of employees to send print requests to specific printers. It can also reroute print requests, if there is a fault or incident, to the nearest appropriate and available printer.

Application- Application servers are designed to install, operate and monitor the applications shared by other devices on the network. Although all users share some applications, such as word processors and email, the server also manages security of restricted access applications such as financial and human resource software.

Database- Some application servers may have a specialised role as database servers, which manage databases with multiple users. These servers are database-architecture independent- the server will manage all activities, whether it is a relational database, flat file or object-oriented.

Web- A web server can also refer to hardware or the web server software. It is an internet server that reacts to HTTP requests to deliver content and services from a client.

Mail- A mail sever is software or a device which acts as an electronic post office. It recieves emails from users on the network, both in the same domain and from external senders, and forwards it to the recipients, who many be local users or external senders.

Hypervisor- This is software written for a particular processor which manages difficult operating systems or copies of the same operating system on a single computer or computer systems. It manages resources  such as the processor, memory and devices so that each operating system not only recieves the resources it needs but also prevents it clashing with the other operating systems.

Virtualisation is to create a simulated or virtual network, computer or server using software.

Normally the operating system and haerdware are closely linked so that only some applications can run simultaneously to prevent problems such as registry conflicts. To operate additional applications, additional physical servers need to be installed, possibly running only a single application.

Storage virtualisation works in a similar way by separating the physical storage server from the logical storage server. It can take two forms:

• Block-level, which allows a storage area network or SAN to use numerous storage arrays across the network as if they were a single array.
• File-level, which allows network attached storage (NAS) and removes the link between the data items and their physcial locations on a particular file server.

Cloud computing is not virtualisation. The cloud uses existing technology to deliver services, such as security, storage and data manipulation, to users over a network. It is used like a utility- you can increase the size of range of your usage through payment.

3.2 Virtualisation

Hybrid cloud computing is a micture of private cloud, public cloud and locally owned services. As its activities and data requirements chaqnge, so the organisation can adjust its usage more effectively and efficiently.

3.3 Networking Characteristics

Peer to Peer- Each computer is linked to one or more computers directly rather than going through a server. There are several forms of peer to peer networks.

The Star Network- All computers communicate through the central hub. All transmissions recieved from a peripheral on the star are broadcast to all peripherals by the central device. The benefit is that of one peripheral goes down the rest can continue to communicate; however, if the hub goes down the whole network is compromised. The cost of the cabling can also be expensive if the network is extensive.

Bus Network- Advantages

• Attaching a new device is easy, so if one peripheral fails the others are not affected.
• It is cheaper in cabling terms than a star network.
• Transmissions are in both directions.
• Transmission is more likely to be read as it is recieved by all devices relating to the recipient.

Disadvantages

• The difficulty of identifying a faulty peripheral.
• If the cable should break either the whole network may be lost or it may be split into two.
• Adding additional devices can slow the network down significantly.
• The possibility of the transmisions crashing into each other.

Ring Network

Ring networks have one directional communication through a cable. The message is received by each computer in the ring between the sender and receiver.

Ring Network

Each computer in the chain sends the data packet to the next computer until the receiving device receives it. Often the message is boosted or regenerated to maintain signal strength.

Rings use a method known as token passing to prevent problems of multiple messages on the network colliding. A small data packet known as the token is continuously transmitted around the ring. When a device has data to transmit, it has to wait until the token arrives. At that point, the computer checks to see if the token is already being used by another device. If not, it adds its own data and control information such a the address of the receiving device and sends the token on its way. When the notice of safe delivery is received the token is released back to the ring.

Mesh Network

There are two types of mesh network:

• Total mesh topology where every node (connection) in the network is connected directly to all other nodes. If a connection stops working all the traffic on the network can be rerouted around the problem.

Mesh Networks

• Partial mesh topology has some nodes connected to only a small number of nodes whereas others are connected to all the other nodes. This reduces the rerouting options if a connection fails, but it is cheaper than the total mesh because there is less cabling.

Cabled mesh networks are very costly because of their complex cabling requirements but wireless mesh networks are cheaper because they do not need cabling.

Client Server

In the networks identified above, all the devices are independent in that they carry out their own tasks with their own software. They may share data but not applications. Client server networks are different. One device, which may be a computer or a specialist server, provides services such as applications, storage or internet connections to its clients (the other computers in the network). Larger networks may have several servers providing different services.

It is possible to link networks through switches, gateways and routers, to allow for greater flexibility.

The choice for topology will depend upon a number of factors:

• The number of computers to be networked.
• The spread of the network, e.g. rroom, building, city, country or continent.
• The cost of topology, e.g. a bus topolgy is cost-effective in a single room with only a few devices. A star topology with a central hub running an ethernet system is a reasonable approach for serveral devices and servers, as long as they are within a single building or in buildings on a single site.
• The amount of money available.

LAN

An alternative is to se an ethernet network LAN. The cbling allows messages to travel in either direction, unlike token ring networks which only communicate in one direction. Half duplex ethernet LANs allow communication in both directions but only in one direction at a time. However, if two messages are simultaneously placed on the network and crash into each other, they may bring the network down.

To enable the netowkr to recover from this problem, a media access control (MAC) method called Carrier Sense Multiple Access with Collision Detection (CSMA/CD) is used. Back off is the randomly calculated time that a device will wait to try and resend the message (the time it will backoff from transmitting). Each device carries out its own calculation and so, in theory, only one will be ready to transmit once the collison is cleared.

WAN And MAN

A wide area network (WAN) can be transglobal, or cover whole countries or continents. Networks operating over small distances such as a town,city or a cluster of close cities are called metropolitan area networks (MAN). Each type of network has its own rules or protocols to ensure that generic devices can communicate.

Examples include ADSL (asymmetric digital suscriber line), which supports faster transmission of data over the copper wire telephone lines tan the traditional voice modem. IDSN (integrated services digital network) is a set of standards for digital transmssion of data, video and voice over traditional telephone copper wire. Leased lines, rather than operating through national and international telephone wires, provide a fixed-bandwidth data connection. This means that the bandwidth is only available to the lease, avoiding a slowdown in transmission when the network is busy.

Voice

Data and voice communication can be sent over telephone wires or by celluar or satellite technology.The clearest and most reliable method is over fixed cables between the sender and reciever. If this is not possible to be linked to a handset or computer which is plugged into a public switched telephone network (PSTN), cellular and satellite communication is appropriate.

However, cellular technology requires the towers which carry transmitter/receivers to be in line if sight to the user's communication device. Disappearing into tunnels on a trainn, standing in a remote field or being shielded by large buildins can disrupt this point to point communication. Also, providers do not like to share their communication systems with other providers as they risk losing revenue from new customers.

Satellites

Satellites in geostatic orbit (those that appear to stay in a single location over earth) are point to multipoint communication systems as they recieve transmissions and rebroadcast them to receivers. If required, the satellite can be moved to another location relatively quickly. However, the distance between the stellite and the ground means there is a delay between the transmission of the signal and its receipt, which can be seen on television news reports from remote locations. Data transmission is also compromised by this issue, especially when trying to transfer large files.

Businesses have developed many systems to manage and manipulate data and aid business practices.

MIS

Management information systems refers to the software and hardware used by managers to obtain the necessary information for accurate decision-making and monitoring the effectiveness of decisions. The system continuously gathers internal and external data, and refines and organises it in one or more databases where it can be interrogated by those with access rights. The data can be downloaded, for example for:

• Reports and other documents.
• Spreadsheets and strategic decision support systems.
• Functional systems such as those for marketing, sales and production.

Used wisely, an MIS can provide evidence of:

• The status of the organisation

• Areas of improvement, e.g. rising sales.
• Areas of risk, e.g. faling sales, loss of market share.
• Personal or group responsibility for success and failure.

Limitations of MIS systems are:

• Cost of creatig and installing the system.
• Poor or inflexible design which does not meet the current or future needs of the organisation.
• Data that is out of date, incomplete or incorrect because of poor error checking facilities.
• End users who do not use the system properly; ths is often caused by lack of training.
• It does not make judgements; it only provides the information needed for decisions.

CRM

To ensure that customer needs are met, data must be gathered, analyse and shared. Customer relations management (CRM) is a process that tracks how the business interacts with current and potential customers. A computer-based CRM system supports such activities by holding records of client communications, meetings and documents. This information is available to those with access rights.

A CRM system allows a business to centrally hold customer information such as:

• Contact details.
• Customer histories.
• Leads for new customers.
• Leads for new services.

Although CRM is primarily aimed at sales and customer support functions, marketers also use CRM to aid their understanding of current sales, prospective sales and forecasting future trends.

Limitations to CRMs include:

• Software issues, such as limits on data storage size or emails.
• Integration difficulties with other organisational systems.
• Lack of training leading to poor data entry or incorrect data.
• Resistance by sales staff who believe that using digital information rather than face to face meetings with clients reduces their ability to understand subtle communication behaviours.

SOP

Standard operating procedures (SOPs) are detailed step by step guides to how functions should be carried out accurately and in the same way each time. The outcomes from each activity should always be the same, leaving an audit trail incase of disputes or external investigations by regulatory bodies, SOPs should alsso be created for business systems such as those outlined above, to ensure that the data and resulting information remains accurate and useful. Software can aid the production of SOPs.

Limitations to SOPs include:

• Imposing restrictions and details which result in infllexible practice and a lack of innovation.
• Too much time being spent on the admin rather thn doing the job.
• The necessity of updating the SOP to reflect new statutory and regulatory requirements.
• Lack of version control. If updates to the SOP are not recorded with date and a brief statement about the changes, individuals could be working to different versions, which can cause serious problems.

Help Desk

When a user of a large, complex or widely spread system has a technical problem, it is useful to have a knowledgeable person ready to solve it. For straightforward problems, instructions can be given to the user and the issue resolved immediately, but complex problems can be actioned by a technician visiting the loaction and resolving the problem or providing a temporary fix. This is normally managed through agreed levels of support for departments or customers.

Limitations include:

• Cost of setting up an in-house help desk with hardware, software and staffing.
• Cost of buying help des services from an external source (third party).
• Issues with availability of the help desk, for exampe, the need or 24-hour call out; availability during bank holidays.
• Loss of service through breakdown of communication systems.